convolutional-autoencoder.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"convolutional-autoencoder.ipynb","provenance":[{"file_id":"1Hff4QhIpiC8EHM_ofp77zIWvFwpyhSpV","timestamp":1603713483108}],"collapsed_sections":[],"toc_visible":true,"authorship_tag":"ABX9TyOnW8TI1djtqizVURjiTki/"},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"markdown","metadata":{"id":"eZ5vLdtnHUmn"},"source":["#### Copyright 2020 Connaissance Publishing."]},{"cell_type":"code","metadata":{"id":"e0itxBa9HWgS"},"source":["# Licensed under the Apache License, Version 2.0 (the \"License\");\n","# you may not use this file except in compliance with the License.\n","# You may obtain a copy of the License at\n","#\n","# https://www.apache.org/licenses/LICENSE-2.0\n","#\n","# Unless required by applicable law or agreed to in writing, software\n","# distributed under the License is distributed on an \"AS IS\" BASIS,\n","# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n","# See the License for the specific language governing permissions and\n","# limitations under the License."],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"iTY6J8glmfS8"},"source":["# **Understanding Deep Learning: Application in Rare Event Prediction**\n","\n","Author: Chitta Ranjan, Ph.D.\n","\n","[ProcessMiner, Inc.](https://processminer.com)\n","\n","cranjan@processminer.com\n","\n","LinkedIn: https://www.linkedin.com/in/chitta-ranjan-b0851911/\n","\n","Medium: https://medium.com/@cran2367\n","\n","\n","---\n"]},{"cell_type":"markdown","metadata":{"id":"J8OGrbgumiTV"},"source":["# Convolutional Autoencoders"]},{"cell_type":"markdown","metadata":{"id":"Nd3wuEGWHkyp"},"source":["In this notebook we will develop models for a rare event prediction problem. The problem used as an example here is taken from a paper manufacturing plant. \n","\n","Link to this notebook: [shorturl.at/krxD9](https://shorturl.at/krxD9)\n","\n","## Problem\n","\n","Paper manufacturing is a continuous manufacturing process. Any unwanted break in the process causes significant losses.\n","\n","\n","![Paper Machine.](https://i.ibb.co/Qk301Sz/paper-machine.png)\n","\n","Paper sheets sometimes breaks in the process due to its fragile nature. Think about what happens when paper tears inside a printer. We have to open the printer, take out the stuck paper, and put the printer back together. This takes time and effort. A sheet break in paper manufacturing is similar but takes several times the time and effort.\n","\n","## Motivation\n","\n","A typical paper manufacturing plant faces at least one sheet break a day. Each such break can cost upwards of \\$25,000. Sheet breaks are known to cost \\$6-8 Million a year in a plant. Considering the worldwide paper industries, this translates to loss revenue of $3 billions every year. \n","\n","Additionally, parts of paper machines have hazardous conditions. Resuming the operations after a sheet break requires operators to face such hazards.\n","\n","Developing a solution to address this problem can save costs and reduce work hazards.\n","\n","## Solution\n","\n","Predicting a sheet break in advance can help in preventing the break. Therefore, here **we will work on developing a prediction model that does early prediction for sheet breaks**.\n","\n","## Challenges\n","\n","* Early prediction--low predictive signal.\n","* Rare event--extremely unbalanced data.\n","* High-dimensional timeseries.\n","\n","----------------"]},{"cell_type":"markdown","metadata":{"id":"ElBWj90KHqFo"},"source":["# Connection"]},{"cell_type":"code","metadata":{"id":"EC_UZmPqHmB4","colab":{"base_uri":"https://localhost:8080/","height":34},"executionInfo":{"status":"ok","timestamp":1604169655625,"user_tz":240,"elapsed":32583,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"89c6a0ed-4f37-4f1c-8b26-8781efb12f03"},"source":["from google.colab import drive\n","drive.mount('/content/gdrive', force_remount=True)\n","root_path = 'gdrive/My Drive/understanding-deep-learning/codebase'\n","import os\n","os.chdir(root_path)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Mounted at /content/gdrive\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"TVMhD2fRH34v"},"source":["os.getcwd()"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"l6MvUvGgIAxy"},"source":["# Libraries"]},{"cell_type":"code","metadata":{"id":"_m8snzeAH9Vl"},"source":["%matplotlib inline\n","\n","import pandas as pd\n","import numpy as np\n","\n","%tensorflow_version 2.x\n","import tensorflow as tf\n","from tensorflow.keras import optimizers\n","from tensorflow.keras.models import Model\n","from tensorflow.keras.models import Sequential\n","from tensorflow.keras.layers import Input \n","from tensorflow.keras.layers import Dense\n","from tensorflow.keras.layers import Dropout\n","from tensorflow.keras.layers import LSTM\n","from tensorflow.keras.layers import RepeatVector\n","from tensorflow.keras.layers import Conv1D\n","from tensorflow.keras.layers import MaxPool1D\n","from tensorflow.keras.layers import Conv1DTranspose\n","from tensorflow.keras.layers import BatchNormalization\n","from tensorflow.keras.layers import Reshape\n","from tensorflow.keras.layers import TimeDistributed\n","from tensorflow.keras.layers import Flatten\n","from tensorflow.keras.layers import Bidirectional\n","from tensorflow.keras.callbacks import ModelCheckpoint\n","from tensorflow.keras.callbacks import TensorBoard\n","from tensorflow.python.keras import backend as K\n","\n","from sklearn.preprocessing import StandardScaler\n","from sklearn.preprocessing import MinMaxScaler\n","from sklearn.model_selection import train_test_split\n","from sklearn.metrics import confusion_matrix\n","from sklearn.metrics import precision_recall_curve\n","from sklearn.metrics import recall_score\n","from sklearn.metrics import classification_report\n","from sklearn.metrics import auc\n","from sklearn.metrics import roc_curve\n","from sklearn.metrics import precision_recall_fscore_support\n","from sklearn.metrics import f1_score\n","\n","from imblearn.over_sampling import SMOTE\n","from collections import Counter\n","\n","import matplotlib.pyplot as plt\n","import seaborn as sns\n","\n","# user-defined libraries\n","import utilities.datapreprocessing as dp\n","import utilities.performancemetrics as pm\n","import utilities.simpleplots as sp\n","import utilities.reconstructionperformance as rp\n","\n","from numpy.random import seed\n","seed(1)\n","\n","SEED = 123  # used to help randomly select the data points\n","DATA_SPLIT_PCT = 0.2\n","\n","from pylab import rcParams\n","rcParams['figure.figsize'] = 6, 4\n","plt.rcParams.update({'font.size': 22})"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"Dm2QJeEXIljC","colab":{"base_uri":"https://localhost:8080/","height":37},"executionInfo":{"status":"ok","timestamp":1604169673845,"user_tz":240,"elapsed":286,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"7ed2b9a4-c236-44bd-ccfe-228e80209043"},"source":["# Check the tensorflow version.\n","tf.__version__"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"application/vnd.google.colaboratory.intrinsic+json":{"type":"string"},"text/plain":["'2.3.0'"]},"metadata":{"tags":[]},"execution_count":5}]},{"cell_type":"markdown","metadata":{"id":"gXYyvWCGmsim"},"source":["# Data Preparation"]},{"cell_type":"markdown","metadata":{"id":"E2XteM5KLLiV"},"source":["## Read and process data"]},{"cell_type":"markdown","metadata":{"id":"vfAum_oRLWWU"},"source":["The data is taken from https://arxiv.org/abs/1809.10717. Please use this source for any citation."]},{"cell_type":"code","metadata":{"id":"ZM-RR7qcIr5K","colab":{"base_uri":"https://localhost:8080/","height":343},"executionInfo":{"status":"ok","timestamp":1604169676870,"user_tz":240,"elapsed":1278,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"5b51f35a-5ef7-466d-ae20-6a6ac9d10559"},"source":["'''Download data here:\n","https://docs.google.com/forms/d/e/1FAIpQLSdPCkTFW1wTp6tAP1C1BhT7HWTQ60-GfCGvWs9Ca4WbR0FIdg/viewform\n","'''\n","df = pd.read_csv(\"data/processminer-sheet-break-rare-event-dataset.csv\")\n","df.head(n=5)  # visualize the data."],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>DateTime</th>\n","      <th>SheetBreak</th>\n","      <th>RSashScanAvg</th>\n","      <th>CT#1 BLADE PSI</th>\n","      <th>P4 CT#2 BLADE PSI</th>\n","      <th>Bleached GWD Flow</th>\n","      <th>ShwerTemp</th>\n","      <th>BlndStckFloTPD</th>\n","      <th>C1 BW SPREAD CD</th>\n","      <th>RS BW SPREAD CD</th>\n","      <th>C1 BW SPREAD MD</th>\n","      <th>RS BW SPREAD MD</th>\n","      <th>C1 BW SCAN AVG</th>\n","      <th>RS BW SCAN AVG</th>\n","      <th>CoatBrkFlo</th>\n","      <th>Clay Flow</th>\n","      <th>CouchLoVac</th>\n","      <th>COUCH VAC</th>\n","      <th>4PrsTopLd</th>\n","      <th>4PrsBotLod</th>\n","      <th>CalndrDrw</th>\n","      <th>2DryrDrw</th>\n","      <th>3DryrDrw</th>\n","      <th>4DryrDraw</th>\n","      <th>1PrsTopDrw</th>\n","      <th>4PrsBotDrw</th>\n","      <th>FanPmpSpd</th>\n","      <th>FlBxHdrVac</th>\n","      <th>FlatBxVac</th>\n","      <th>Grade&amp;Bwt</th>\n","      <th>UnblGWDFlo</th>\n","      <th>Hdbox pH</th>\n","      <th>HdBxLiqLvl</th>\n","      <th>TotHead\"</th>\n","      <th>HorzSlcPos</th>\n","      <th>KraftFlow</th>\n","      <th>CouchLoad</th>\n","      <th>C1MoSprdCD</th>\n","      <th>RSMoSprdCD</th>\n","      <th>C1MoSprdMD</th>\n","      <th>RSMoSprdMD</th>\n","      <th>RL MoisAct</th>\n","      <th>PrScrRjFlo</th>\n","      <th>RwBrkFlo</th>\n","      <th>RcycFbrFlo</th>\n","      <th>RetnAidFlo</th>\n","      <th>RUSH DRAG</th>\n","      <th>Rush/Drag</th>\n","      <th>SilicaFlo</th>\n","      <th>HBxSlcTemp</th>\n","      <th>SodAlumFlo</th>\n","      <th>CouchSpd</th>\n","      <th>MachSpd</th>\n","      <th>1PrsTopSpd</th>\n","      <th>4PrsBotSpd</th>\n","      <th>WtNStarFlo</th>\n","      <th>BasWgtFlo</th>\n","      <th>TMP Flow</th>\n","      <th>HBxTotHead</th>\n","      <th>TrayCons</th>\n","      <th>UpprHdTmpRL</th>\n","      <th>VertSlcPos</th>\n","      <th>EventPress</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>5/1/99 0:00</td>\n","      <td>0</td>\n","      <td>0.376665</td>\n","      <td>-4.596435</td>\n","      <td>-4.095756</td>\n","      <td>13.497687</td>\n","      <td>-0.118830</td>\n","      <td>-20.669883</td>\n","      <td>0.000732</td>\n","      <td>-0.061114</td>\n","      <td>-0.059966</td>\n","      <td>-0.038189</td>\n","      <td>0.877951</td>\n","      <td>-0.052959</td>\n","      <td>-13.306135</td>\n","      <td>0.101068</td>\n","      <td>0.041800</td>\n","      <td>0.199901</td>\n","      <td>-2.327329</td>\n","      <td>-0.944167</td>\n","      <td>3.075199</td>\n","      <td>0.123154</td>\n","      <td>-0.104334</td>\n","      <td>-0.570710</td>\n","      <td>-9.784456</td>\n","      <td>0.355960</td>\n","      <td>15.842819</td>\n","      <td>-0.451974</td>\n","      <td>-0.105282</td>\n","      <td>96</td>\n","      <td>-134.277862</td>\n","      <td>0.058726</td>\n","      <td>-0.021645</td>\n","      <td>9.366755</td>\n","      <td>0.002151</td>\n","      <td>-69.187583</td>\n","      <td>4.232571</td>\n","      <td>-0.225267</td>\n","      <td>-0.196872</td>\n","      <td>-0.072449</td>\n","      <td>-0.103732</td>\n","      <td>-0.720746</td>\n","      <td>-5.412436</td>\n","      <td>76.679042</td>\n","      <td>-0.632727</td>\n","      <td>1351.632859</td>\n","      <td>-0.657095</td>\n","      <td>-0.434947</td>\n","      <td>-108.775969</td>\n","      <td>0.084856</td>\n","      <td>10.210182</td>\n","      <td>11.295155</td>\n","      <td>29.984624</td>\n","      <td>10.091721</td>\n","      <td>0.053279</td>\n","      <td>-4.936434</td>\n","      <td>-24.590146</td>\n","      <td>18.515436</td>\n","      <td>3.473400</td>\n","      <td>0.033444</td>\n","      <td>0.953219</td>\n","      <td>0.006076</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>5/1/99 0:02</td>\n","      <td>0</td>\n","      <td>0.475720</td>\n","      <td>-4.542502</td>\n","      <td>-4.018359</td>\n","      <td>16.230658</td>\n","      <td>-0.128733</td>\n","      <td>-18.758079</td>\n","      <td>0.000732</td>\n","      <td>-0.061114</td>\n","      <td>-0.059966</td>\n","      <td>-0.038189</td>\n","      <td>0.873273</td>\n","      <td>-0.014244</td>\n","      <td>-13.306135</td>\n","      <td>0.101108</td>\n","      <td>0.041447</td>\n","      <td>0.304313</td>\n","      <td>-2.340627</td>\n","      <td>-0.939994</td>\n","      <td>3.075199</td>\n","      <td>0.123154</td>\n","      <td>-0.104334</td>\n","      <td>-0.574861</td>\n","      <td>-9.784456</td>\n","      <td>0.360160</td>\n","      <td>16.491684</td>\n","      <td>-0.450451</td>\n","      <td>-0.092430</td>\n","      <td>96</td>\n","      <td>-134.480194</td>\n","      <td>0.058759</td>\n","      <td>-0.004579</td>\n","      <td>9.350215</td>\n","      <td>0.002149</td>\n","      <td>-68.585197</td>\n","      <td>4.311490</td>\n","      <td>-0.225267</td>\n","      <td>-0.196872</td>\n","      <td>-0.059103</td>\n","      <td>-0.083895</td>\n","      <td>-0.720746</td>\n","      <td>-8.343222</td>\n","      <td>78.181598</td>\n","      <td>-0.632727</td>\n","      <td>1370.378949</td>\n","      <td>-0.875628</td>\n","      <td>-1.125819</td>\n","      <td>-108.848967</td>\n","      <td>0.085146</td>\n","      <td>12.534340</td>\n","      <td>11.290761</td>\n","      <td>29.984624</td>\n","      <td>10.095871</td>\n","      <td>0.062801</td>\n","      <td>-4.937179</td>\n","      <td>-32.413266</td>\n","      <td>22.760065</td>\n","      <td>2.682933</td>\n","      <td>0.033536</td>\n","      <td>1.090502</td>\n","      <td>0.006083</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>5/1/99 0:04</td>\n","      <td>0</td>\n","      <td>0.363848</td>\n","      <td>-4.681394</td>\n","      <td>-4.353147</td>\n","      <td>14.127997</td>\n","      <td>-0.138636</td>\n","      <td>-17.836632</td>\n","      <td>0.010803</td>\n","      <td>-0.061114</td>\n","      <td>-0.030057</td>\n","      <td>-0.018352</td>\n","      <td>1.004910</td>\n","      <td>0.065150</td>\n","      <td>-9.619596</td>\n","      <td>0.101148</td>\n","      <td>0.041095</td>\n","      <td>0.252839</td>\n","      <td>-2.353925</td>\n","      <td>-0.935824</td>\n","      <td>3.075199</td>\n","      <td>0.123154</td>\n","      <td>-0.104334</td>\n","      <td>-0.579013</td>\n","      <td>-9.784456</td>\n","      <td>0.364356</td>\n","      <td>15.972885</td>\n","      <td>-0.448927</td>\n","      <td>-0.097144</td>\n","      <td>96</td>\n","      <td>-133.946587</td>\n","      <td>0.058791</td>\n","      <td>-0.084658</td>\n","      <td>9.037409</td>\n","      <td>0.002148</td>\n","      <td>-67.838187</td>\n","      <td>4.809914</td>\n","      <td>-0.225267</td>\n","      <td>-0.186801</td>\n","      <td>-0.048696</td>\n","      <td>-0.073823</td>\n","      <td>-0.720746</td>\n","      <td>-1.085166</td>\n","      <td>79.684154</td>\n","      <td>-0.632727</td>\n","      <td>1368.123089</td>\n","      <td>-0.037775</td>\n","      <td>-0.519541</td>\n","      <td>-109.086577</td>\n","      <td>0.085436</td>\n","      <td>18.582893</td>\n","      <td>11.286366</td>\n","      <td>29.984624</td>\n","      <td>10.100265</td>\n","      <td>0.072322</td>\n","      <td>-4.937924</td>\n","      <td>-34.183774</td>\n","      <td>27.004663</td>\n","      <td>3.537487</td>\n","      <td>0.033629</td>\n","      <td>1.840540</td>\n","      <td>0.006090</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>5/1/99 0:06</td>\n","      <td>0</td>\n","      <td>0.301590</td>\n","      <td>-4.758934</td>\n","      <td>-4.023612</td>\n","      <td>13.161566</td>\n","      <td>-0.148142</td>\n","      <td>-18.517601</td>\n","      <td>0.002075</td>\n","      <td>-0.061114</td>\n","      <td>-0.019986</td>\n","      <td>-0.008280</td>\n","      <td>0.930037</td>\n","      <td>-0.067199</td>\n","      <td>-15.196531</td>\n","      <td>0.101188</td>\n","      <td>0.040742</td>\n","      <td>0.072873</td>\n","      <td>-2.367223</td>\n","      <td>-0.931651</td>\n","      <td>3.075199</td>\n","      <td>0.123154</td>\n","      <td>-0.104334</td>\n","      <td>-0.583165</td>\n","      <td>-9.784456</td>\n","      <td>0.368556</td>\n","      <td>15.608688</td>\n","      <td>-0.447404</td>\n","      <td>-0.160073</td>\n","      <td>96</td>\n","      <td>-134.002586</td>\n","      <td>0.058824</td>\n","      <td>-0.055118</td>\n","      <td>9.020625</td>\n","      <td>0.002146</td>\n","      <td>-67.091148</td>\n","      <td>5.308343</td>\n","      <td>-0.225267</td>\n","      <td>-0.186801</td>\n","      <td>-0.047017</td>\n","      <td>-0.063752</td>\n","      <td>-0.720746</td>\n","      <td>6.172891</td>\n","      <td>81.186702</td>\n","      <td>-0.632727</td>\n","      <td>1365.691449</td>\n","      <td>-0.987410</td>\n","      <td>0.674524</td>\n","      <td>-109.562774</td>\n","      <td>0.085726</td>\n","      <td>17.719032</td>\n","      <td>11.281972</td>\n","      <td>29.984624</td>\n","      <td>10.104660</td>\n","      <td>0.081600</td>\n","      <td>-4.938669</td>\n","      <td>-35.954281</td>\n","      <td>21.672449</td>\n","      <td>3.986095</td>\n","      <td>0.033721</td>\n","      <td>2.554880</td>\n","      <td>0.006097</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>5/1/99 0:08</td>\n","      <td>0</td>\n","      <td>0.265578</td>\n","      <td>-4.749928</td>\n","      <td>-4.333150</td>\n","      <td>15.267340</td>\n","      <td>-0.155314</td>\n","      <td>-17.505913</td>\n","      <td>0.000732</td>\n","      <td>-0.061114</td>\n","      <td>-0.030057</td>\n","      <td>-0.008280</td>\n","      <td>0.828410</td>\n","      <td>-0.018472</td>\n","      <td>-14.609266</td>\n","      <td>0.101229</td>\n","      <td>0.040390</td>\n","      <td>0.171033</td>\n","      <td>-2.380521</td>\n","      <td>-0.927478</td>\n","      <td>3.075199</td>\n","      <td>0.123154</td>\n","      <td>-0.104334</td>\n","      <td>-0.587316</td>\n","      <td>-9.784456</td>\n","      <td>0.372756</td>\n","      <td>15.606125</td>\n","      <td>-0.445880</td>\n","      <td>-0.131630</td>\n","      <td>96</td>\n","      <td>-133.145706</td>\n","      <td>0.058856</td>\n","      <td>-0.153851</td>\n","      <td>9.344233</td>\n","      <td>0.002145</td>\n","      <td>-65.991813</td>\n","      <td>5.806771</td>\n","      <td>-0.225267</td>\n","      <td>-0.186801</td>\n","      <td>-0.057088</td>\n","      <td>-0.063752</td>\n","      <td>-0.720746</td>\n","      <td>-3.379599</td>\n","      <td>82.689258</td>\n","      <td>-0.632727</td>\n","      <td>1363.257859</td>\n","      <td>-0.238444</td>\n","      <td>-0.063044</td>\n","      <td>-110.038909</td>\n","      <td>0.086016</td>\n","      <td>16.855202</td>\n","      <td>11.277577</td>\n","      <td>29.984624</td>\n","      <td>10.109054</td>\n","      <td>0.091121</td>\n","      <td>-4.939414</td>\n","      <td>-37.724789</td>\n","      <td>21.907251</td>\n","      <td>3.601573</td>\n","      <td>0.033777</td>\n","      <td>1.410494</td>\n","      <td>0.006105</td>\n","      <td>0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["      DateTime  SheetBreak  RSashScanAvg  ...  UpprHdTmpRL  VertSlcPos  EventPress\n","0  5/1/99 0:00           0      0.376665  ...     0.953219    0.006076           0\n","1  5/1/99 0:02           0      0.475720  ...     1.090502    0.006083           0\n","2  5/1/99 0:04           0      0.363848  ...     1.840540    0.006090           0\n","3  5/1/99 0:06           0      0.301590  ...     2.554880    0.006097           0\n","4  5/1/99 0:08           0      0.265578  ...     1.410494    0.006105           0\n","\n","[5 rows x 63 columns]"]},"metadata":{"tags":[]},"execution_count":6}]},{"cell_type":"code","metadata":{"id":"TpXyxgnkLZhB"},"source":["# Convert Categorical column to hot dummy columns\n","hotencoding1 = pd.get_dummies(df['Grade&Bwt'])\n","hotencoding1 = hotencoding1.add_prefix('grade_')\n","hotencoding2 = pd.get_dummies(df['EventPress'])\n","hotencoding2 = hotencoding2.add_prefix('eventpress_')\n","\n","df=df.drop(['Grade&Bwt', 'EventPress'], axis=1)\n","\n","df=pd.concat([df, hotencoding1, hotencoding2], axis=1)\n","\n","# Rename response column name for ease of understanding\n","df=df.rename(columns={'SheetBreak':'y'})\n","\n","# Shift the response column y by 2 rows to do a 4-min ahead prediction.\n","df = dp.curve_shift(df, shift_by=-2)\n","\n","# Sort by time and drop the time column.\n","df['DateTime'] = pd.to_datetime(df.DateTime)\n","df = df.sort_values(by='DateTime')\n","df = df.drop(['DateTime'], axis=1)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"7GGllJp3vQVp"},"source":["input_X = df.loc[:, df.columns != 'y'].values  # converts df to numpy array\n","input_y = df['y'].values"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"oHOIMk0FL-EN"},"source":["\n","## Temporalization \n","\n","Temporal models, such as LSTM and CNN, requires temporalization. Data temporalization means creating block sequences where each block is a time window. In temporalization, each block is mapped with a corresponding response label. In temporalization process, we will create 3-dimensional arrays of shape: `(samples, timesteps, features)`. Samples is the number of data points. Timesteps is the size of a block, i.e. the time window. This is sometimes called the lookback period. The features is the number of features in a sample."]},{"cell_type":"code","metadata":{"id":"O6War7v0Luhg"},"source":["# Temporalize the data\n","lookback = 20\n","X, y = dp.temporalize(X=input_X, y=input_y, lookback=lookback)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"l6EpvEC4Mw9N"},"source":["## Data split into train, valid, and test"]},{"cell_type":"code","metadata":{"id":"s24c7L1nMDwF"},"source":["X_train, X_test, y_train, y_test = train_test_split(np.array(X),\n","                                                    np.array(y),\n","                                                    test_size=DATA_SPLIT_PCT,\n","                                                    random_state=SEED)\n","X_train, X_valid, y_train, y_valid = train_test_split(X_train,\n","                                                      y_train,\n","                                                      test_size=DATA_SPLIT_PCT,\n","                                                      random_state=SEED)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"O_El1abiM0v4","colab":{"base_uri":"https://localhost:8080/","height":34},"executionInfo":{"status":"ok","timestamp":1604169697677,"user_tz":240,"elapsed":264,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"a33d1825-593f-474d-f4b7-b991a32ef2a9"},"source":["X_train.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(12992, 20, 69)"]},"metadata":{"tags":[]},"execution_count":11}]},{"cell_type":"markdown","metadata":{"id":"vE03z36ZM7Ym"},"source":["# Rare Event Prediction as Anomaly Detection\n","\n","If an event is extremely rare, we can use Anomaly Detection approach to predict the rare event. In this approach, the rare event is treated as anomalies. The model is trained to detect the anomaly. \n","\n","The Anomaly Detection approach works using a Reconstruction Model. The approach is:\n","\n","- divide the data into two parts: majority class (negatively labeled) and minority class (positively labeled).\n","\n","- The majority class is treated as *normal* state of the process. A normal state is when the process is eventless.\n","\n","- Train a reconstruction model on the normal state samples (i.e. ignore the positively labeled data).\n","\n","- A well-trained reconstruction model will be able to accurately reconstruct a new sample if it belongs to the **normal state**, i.e. have a **small reconstruction error**.\n","\n","- However, a sample from a **rare-event** would be *abnormal* for the model. The reconstruction model will struggle reconstructing it, i.e. we will have **large reconstruction error**.\n","\n","- We catch such high reconstruction error instances and label them as a rare event prediction.\n","\n","# Challenges\n","There are some critical challenges in this approach that we need to keep into considerations. High-dimensional time-series, such as the sheet break problem here, can be both,\n","\n","a. non-stationary, and\n","\n","b. contain outliers.\n","\n","Non-stationarity can be dealt with by doing an online model training, i.e. periodically retrain the model with the most recent window of data. This works under the assumption that the process is stationary within the window. This approach is recommended to be followed when deploying this system for real-time operations. Here we will skip it.\n","\n","Handling outliers could be more challenging. Identifying outliers in a high-dimensional time-series is non-trivial. One possible approach is using PCA and detecting outliers in a reduced dimension of the PC space. Or we can train a reconstruction model, remove the training samples that have high reconstruction error, and iterate until we do not have such training samples."]},{"cell_type":"markdown","metadata":{"id":"p7Ckx0euNG-Y"},"source":["# Extract ''Normal'' samples\n","\n","As part of the anomaly detection approach, we require to extract the samples that belong to the ''normal'' state, i.e. the samples with response as 1."]},{"cell_type":"code","metadata":{"id":"E2Wta6liM2zA"},"source":["X_train_y0 = X_train[y_train==0]  # Normal state samples.\n","X_train_y1 = X_train[y_train==1]  # Anomalous samples.\n","\n","X_valid_y0 = X_valid[y_valid==0]  # Normal state samples.\n","X_valid_y1 = X_valid[y_valid==1]  # Anomalous samples."],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"nUYK8zsqNJ6z","colab":{"base_uri":"https://localhost:8080/","height":34},"executionInfo":{"status":"ok","timestamp":1604169700679,"user_tz":240,"elapsed":216,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"13839024-d89f-4dc5-c4f5-bde1f8844fd5"},"source":["X_train_y0.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(12818, 20, 69)"]},"metadata":{"tags":[]},"execution_count":13}]},{"cell_type":"code","metadata":{"id":"rudYh1HANMXI"},"source":["# Initialize a scaler using the training data.\n","scaler = StandardScaler().fit(dp.flatten(X_train_y0))\n","\n","X_train_y0_scaled = dp.scale(X_train_y0, scaler)\n","X_train_y1_scaled = dp.scale(X_train_y1, scaler)\n","X_train_scaled = dp.scale(X_train, scaler)\n","\n","X_valid_scaled = dp.scale(X_valid, scaler)\n","X_valid_y0_scaled = dp.scale(X_valid_y0, scaler)\n","\n","X_test_scaled = dp.scale(X_test, scaler)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"uj-LI7sXNObk"},"source":["TIMESTEPS =  X_train_y0_scaled.shape[1]\n","N_FEATURES =  X_train_y0_scaled.shape[2]"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"xEuXNTXdlnND"},"source":["# Sparse Convolutional Autoencoder"]},{"cell_type":"code","metadata":{"id":"JRy3rgWpowiU","colab":{"base_uri":"https://localhost:8080/","height":1000},"executionInfo":{"status":"ok","timestamp":1604169705816,"user_tz":240,"elapsed":1233,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"9426b2af-d216-4028-c099-3b9364488d0c"},"source":["inputs = Input(shape=(TIMESTEPS, \n","                      N_FEATURES), \n","               name='encoder_input')\n","\n","def encoder(inp):\n","  '''\n","  Encoder.\n","\n","  Input\n","  inp   A tensor of input data.\n","\n","  Process\n","  Extract the essential features of the input as\n","  its encodings by filtering it through convolutional\n","  layer(s). Pooling can also be used to further\n","  summarize the features.\n","\n","  A linearly activated dense layer is added as the\n","  final layer in encoding to perform any affine\n","  transformation required. The dense layer is not\n","  for any feature extraction. Instead, it is only\n","  to make the encoding and decoding connections\n","  simpler for training.\n","\n","  Output\n","  encoding  A tensor of encodings.\n","  '''\n","\n","  # Multiple (conv, pool) blocks can be added here\n","  conv1 = Conv1D(filters=N_FEATURES,\n","                 kernel_size=4,\n","                 activation='relu',\n","                 padding='same',\n","                 name='encoder-conv1')(inp)\n","  pool1 = MaxPool1D(pool_size=4,\n","                    strides=1,\n","                    padding='same',\n","                    name='encoder-pool1')(conv1)\n","  \n","  # The last layer in encoding\n","  encoding = Dense(units=N_FEATURES,\n","                   activation='linear',\n","                   activity_regularizer=\n","                    tf.keras.regularizers.L1(l1=0.01),\n","                   name='encoder')(pool1)\n","  \n","  return encoding\n","\n","def decoder(encoding):\n","  '''\n","  Decoder.\n","\n","  Input\n","  encoding    The encoded data.\n","\n","  Process\n","  The decoding process requires a transposed \n","  convolutional layer, a.k.a. a deconvolution \n","  layer. Decoding must not be done with a \n","  regular convolutional layer. A regular conv \n","  layer is meant to extract a downsampled \n","  feature map. Decoding, on the other hand, \n","  is reconstruction of the original data from \n","  the downsampled feature map. A regular \n","  convolutional layer would try to extract \n","  further higher level features from\n","  the encodings instead of a reconstruction.\n","  \n","  For a similar reason, pooling must not be \n","  used in a decoder. A pooling operation is \n","  for summarizing a data into a few summary\n","  statistics which is useful in tasks such as \n","  classification. The purpose of decoding is \n","  the opposite, i.e., reconstruct the original \n","  data from the summarizations. Adding pooling \n","  in a decoder makes it lose the variations \n","  in the data and, hence, a poor reconstruction.\n","\n","  If the purpose is only reconstruction, a \n","  linear activation should be used in decoding. \n","  A nonlinear activation is useful for\n","  predictive features but not for reconstruction.\n","\n","  Batch normalization helps a decoder by \n","  preventing the reconstructions\n","  from exploding.\n","\n","  Output\n","  decoding    The decoded data.\n","\n","  '''\n","\n","  convT1 = Conv1DTranspose(filters=N_FEATURES,\n","                           kernel_size=4,\n","                           activation='linear',\n","                           padding='same',\n","                           name='decoder-convT1')(encoding)\n","\n","  decoding = BatchNormalization(\n","      name='decoder-batchnorm'\n","  )(convT1)\n","\n","  decoding = Dense(units=N_FEATURES,\n","                   activation='linear',\n","                   name='decoder')(decoding)\n","  \n","  return decoding\n","\n","autoencoder = Model(inputs=inputs, \n","                    outputs=decoder(encoder(inputs)))\n","\n","autoencoder.summary()\n","\n","# Plot the network structure\n","tf.keras.utils.plot_model(model=autoencoder, \n","                          show_shapes=False, \n","                          dpi=160)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Model: \"functional_1\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","encoder_input (InputLayer)   [(None, 20, 69)]          0         \n","_________________________________________________________________\n","encoder-conv1 (Conv1D)       (None, 20, 69)            19113     \n","_________________________________________________________________\n","encoder-pool1 (MaxPooling1D) (None, 20, 69)            0         \n","_________________________________________________________________\n","encoder (Dense)              (None, 20, 69)            4830      \n","_________________________________________________________________\n","decoder-convT1 (Conv1DTransp (None, 20, 69)            19113     \n","_________________________________________________________________\n","decoder-batchnorm (BatchNorm (None, 20, 69)            276       \n","_________________________________________________________________\n","decoder (Dense)              (None, 20, 69)            4830      \n","=================================================================\n","Total params: 48,162\n","Trainable params: 48,024\n","Non-trainable params: 138\n","_________________________________________________________________\n"],"name":"stdout"},{"output_type":"execute_result","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAj0AAAQxCAIAAABK4asXAAAABmJLR0QA/wD/AP+gvaeTAAAgAElEQVR4nOzdeUBU5f4/8GeAmWHGkVUQUSAdJUgIwcgFARdcUEwEESG1m3m7KS5o99u3fqXeMssyUcvymmm5kF7QsszcUkQlF1JxQc2McCXZQYZhmzm/P577PXfuAOMwC8MzvF9/DWeec87nObO8zzlzzoOA4zgCAADACBtLFwAAANAGyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWGJn6QI6OqVSWVZWZukqAKCzkEqlLi4ulq6iQ0NuPcYPP/wwdepUS1cBAJ1FQkJCRkaGpavo0HCeEAAAWILjLb306NHj3Llzlq4CAKzZO++8s2nTJktXwQDkll5sbW179epl6SoAwJrJZDJLl8AGnCcEAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXIL9LJt27aQkBAHBweBQCAQCHbs2GHpigCgk0JugV5mzpx54cKFCxcuWLoQAOjskFvAsJs3b/r7+3t6ev7000+WrqVdzZgx48knn5RIJPTw19bWtlevXsOGDbt+/bqlS2ubsrKy4cOHP/HEE3Z2drQvQqHQy8tr8ODBW7ZssXR10EEht4Bhu3btunHjRlFR0YYNGyxdS7vavn37r7/++u2339I/Z8yYce/evVOnTvn7+1u2sLZydXU9fvx4YWHhW2+9RaesWbPm7t27Z86cmTVrlmVrgw4LuQUMmzp1ar9+/dzd3f/2t79ZuhYAaCd2li4AwHB+fn43b960dBVt89prrxUUFOzevdvShVhGJ+8+mASOt6ANbGzwhjHKw4cPP/30U0tXYTGdvPtgKvgagjYQCoWWLoFtK1eurK2ttXQVFtPJuw+mgtxqDzU1NStXrnz22WednJxEIpG3t/f06dPz8vLos3l5ecHBwUKhUCAQBAcHE0JOnz49Z84cX19fiUQil8vnz5+vUChaW/ixY8dmzJghl8ulUqm9vb2Hh0dAQEBcXFxWVhbfRqVS7dixIyYmxtPTUywWOzo6BgYGLly4sKCgoLXFqlSqrVu3RkdHe3h4iEQiBwcHuVw+c+ZMw/pICMnKygoKCqLdDA0NJYSo1erdu3ePHz/e2dm5pqamLVuUEELi4+M9PT3pRWjz5s3jp7d1e547dy4wMNDW1lYgEMTFxR05ciQhIaFnz54ikcjd3f25555rfrHiZ5991rdvXxsbG4FAEBUVxU/fvXt3SEiIVCqlVV29epV/at26dQEBAevWrSOE7NmzhzZwcnKiz1ZVVQ0fPtzFxWXTpk1t3Q5aWOy+Ma5fvz5lypSgoKAePXpIJBKZTNavX7+XX375wYMHtMHq1avd3d0F/8fT0zM6Opqffdq0ab1797axsbG1tQ0KCtL81JDHvatN/pYGfXGgU0ZGBiGkV69eBi/h9u3bvr6+hJAFCxbcu3dPoVCkp6dLJBKhULh3716+WUpKCiHE0dFx6NChzV+m+Pj45ktuaGhITk4mhAwePPjYsWMKhaK0tPTnn3+mi1q1ahVtVl5eHh4eTghJSkq6evVqXV3dgwcP0tLSpFKpWCxOT09vvuSysrKwsDBCyKhRo86ePatUKpVKZXFx8b59+2g927dvN6CPfDflcvmRI0f69+/Pd/DRo0eGbV5XV1dCSEpKitb0tm7PCRMmEEIkEkmLH5N//OMfzVc9aNAguom0pmdmZtK5rly5ovUU7XLztaelpdFZ7O3t6+vr9ez7gQMH6FwvvPAC093nOG7ZsmV0rk8++UTP7vMbwcbG5qOPPrp27ZpCobh16xaNUh8fn6qqKtqmtrbW2dmZEOLh4dHY2Ki1BLVa/cQTTyQkJGhNb9Mn1yRv6UWLFhFCmlcCWpBbj2F8bg0bNowQMmnSJM2Jb7/9NiHE2dmZ/2jRd79MJsvPzy8pKamrq1MoFIcPH/b09KQfg7t372otec6cOYQQf3//2tparaf69evH51ZsbCwhJCoqSq1Wa7bZuXMnIUQoFObm5mrNTr/FQkJCtL5Df/vtN1qMVm7p2Ue+m0KhcOLEiWvXrj18+DDdBzdTbum/PWmXe/bsuXfv3vv379fX1xcXF2/cuNHBwYG2P3TokNYqTPjF/dFHH9FZxGKxCXOLle5zxuXWiBEjNKf8+uuvdFGbNm3iJ9JIIIRo7UhxHHf06FFCyJEjR7Smt+mTa5K3NHJLTzhPaF4nT548deoUIYRmDC8+Pp4QUlFRsXfvXs3ptra2Tz31VLdu3cRisVQqHT16NP95vnbtmmbL69evb9y4kRDy+uuvN99NjoyMpA/Onz9PV7Fo0SKBQKDZJjExUS6XNzY2/uMf/9CcnpOTs3//fkLIG2+8IRKJTN5HQohUKv3+++8XLlw4evToefPm+fn5PXYthtF/e1IeHh6TJk3y9PQUiURubm4vv/wy3XEhhHzwwQdmKpIQMmvWrIiICCcnp7Vr1+qzzfXESveNMW7cuGPHjmlO6devH/0hlt/NIoTMmTOHvv+bn4ndsmVLnz59Ro0apTmxre/qdntLA8HvW+bG7xEHBgZqTvfx8aEPzp49q3sJ/IxKpVJzekZGhlqtJoSMHj26+VwpKSkjR44khOzZs4dOGTJkiFYbgUBA4+3QoUOaC+f3l1tccnNG9tHGxub69esymUyfdRmvte3ZmrFjxwYFBRFCsrOzzXdNgbOzc3Z2dkVFxSuvvGKmVVAds/umJRAIaPbX19fzE/v160fPHx48ePD+/fv89Orq6m+++Wb27Nlae3XGvKvb+S3dCSG3zCs/P58QIhAI3N3dNafLZDI7OztCSGlpqe4l8JfwcRynOf2XX34hhEgkkh49ejSfa8CAASEhIYSQS5cu0Wb0/L4WuVxOCGlqaqJ1UhcvXiSEuLu7Ozo6Pq5/hJiij+2pte2pAz1fpFKpWjxGYUtn7v7cuXMJISqVSnMEqZ07dzY2Nr744otajdl6V3c2yC3zqqysJIRwHEcvOtLU1NRE/nuXsE1KSkoIIV26dNHdrKysTEcz/tcL2kxzyfxTj2W+PnYQPXv2pA8qKiosW4lFdPDuV1VVrVmzJjo6um/fvu7u7jKZTCgUtnj97cSJE728vAghW7ZsoecqCCFffvnlc8895+HhodXY6t/VTMN4Ge1BIpGY6RxLZWWlWq02+HZg/tOriX4sGxoa2rQo8/XR4vhjFLqj3dl05O7fv38/IiKioKAgLi5uy5YtcrncxcVFKBQ6OTk1jy5bW9uXX355yZIlhYWFP/3005gxY65fv3727Fn+lGBzVvyuZhqOt8yL3qGiVCpNfjOHm5sbIaSpqYkeHrWmW7duhJDW1l5dXa3ZjKKnB4uLi1UqlT6VmK+PHcSjR4/oA61TRp1Ex+l+fX29SCTi6yGELFu2rKCgICAgICMjIyIiomfPnhKJREe+/vWvf6UxTK/O2LJli4+Pz5gxY5q3tPp3NdOQW+bFX1ZEf2cyIXpLKSHkyJEjj21WV1eneSaQd+vWLUKIUCgMCAjgJz755JN0lnPnzulTifn62EHQrdSlS5cWLxLTM93Z1XG6f/ny5cbGRqlUyk+hF0cEBwfb2trqs4Tu3bvHxcURQr7//vuioqIdO3bMnj27xdMVVv+uZhpyy7z4S/JWrFjR4kk5g9HrcQkhS5YsefjwYWvN6KeUEJKTk6P1FMdx2dnZhJCYmBixWMxP50dAeP/99/WpxHx97AhUKhUdQyEuLk7ry5HefvD777/rf40DvWit+S58ZWVlZGSki4sLvbeh42if7uvpl19+sbe31yyjrq6OEPLnn3/qvxB6XXtDQ8Pzzz9fUlLS2n9Lse53NeuQW+Y1cuTIgQMHEkIOHDgwduzYo0eP0l+kqqur8/Pz9+zZo3nSo02CgoISExMJIYWFhaGhoRs3bvzjjz/q6+tVKlVpaem5c+e++eYbQkhwcPDkyZMJIWvWrNH6ftmxY0dhYaFQKFy6dKnm9ISEhF69ehFC9u3bl5SUdPny5YaGBoVCcfv27ZMnT7ZnHzuCrVu3FhUVSSSSJUuWaD1F/9nV3bt3X3vttYcPHzY2NpaUlPz88886/oll9+7dCSHZ2dknTpxoaGgoKiqi+xxbtmw5ceJERUVFampqW39ZNKv26b6ezp8/r3VxeZ8+fQghR48e3b59u0KhqKmpOXny5OzZs3Vc5R8ZGUlvf87KyqIjn7XYzLrf1cyz1A3PrDB+vIybN2/y93xo8fb2Li8vp834gXm0Zs/NzaWNv/32W62nqqurNUda0xIbG0ublZeX00GbEhMTr169Wl9fX1RUlJaWJpFIxGLxzp07m9eck5PT4vWE/Idca7wMPfuoo5sGe+w4T1rTW9uedMAI+o9Rqqurm5qaHj58uG7dOolEIpVKf/jhh+arvnDhQvOfUmxtbelQdaSlASNWrVql1Z7uT6xevZr+adrxMljpPtf6eBmNjY2PHj26c+dOTk6Or6+vj4+P5rP8f87UNHjwYJpnCxcubHG7rV+/nrZssV88Iz+5BsB4GXpCbj2G8bnFcVxVVdV77703dOhQFxcXOzs7Nze3qKioTz/9tKamhuO4ixcvBgcH818Bnp6eL774Isdx9NwRfxOVTCaLjIxsaGjQXLJarf7uu+8SEhJ8fHxoDvn4+MTHx+/atUulUvHNGhsbv/zyyzFjxri5udnZ2Tk4OAQGBqamphYUFLRWc2FhYUpKSt++fe3t7aVS6YABA1asWFFdXU3zTCqVhoWFlZSU6NlHjuOOHTsWFBTEd9PNzS0sLCwrK8uwTZqTkxMWFkYvTqFfl3369Bk3bpxKpTJse9Ivbnd394EDB9Jr0qRSqb+/f2pq6p07d1or44cffhg0aFDXrl2dnJzCw8Pff//9O3funDlzhq7C2dk5Ojpas31dXd38+fO7d+9ub2/v7+8/d+7cwsJCjuMqKiroeBkbNmzQp/t/+ctf/P397e3t6YpsbGy8vLwiIyOvX7/OVvcvXLgwatSo3r1763mx4lNPPaVVw8aNG/38/EQiUbdu3SIiIr744oumpqapU6cSQoRC4YABA06fPq01S3V1tUwm8/Ly0vyMtEj3u9q0b2kOuaU35NZjmCS3oOOjX9wRERGWLsQyOlX3VSpVr169Whws2LKQW3rC71sA/6Hn6EfWqpN0/+DBg0VFRS+99JKlCwEDIbfA8o4fPy5oi+HDh5upkk5+5Vgn6f6HH34YExNDLz4CFnW4G+ABAEzo/v371dXV9OpHQsi2bduys7Pp/y4BRuF4Cyxv+PDhbTq7ffz4cUuXDMzIzc194YUXiouLa2trN2zY8PLLL48ePZr+twRgFI63AMDK5ebm0lvHCCE9evT44osvLFsPGAnHW9DZff31135+fvReqPPnz3t6eoaHh+fl5Vm6rnZi9d13dXXt1auXra2tg4PDlClTzp496+3tbemiwCg43oLOLjk5OTk52dJVWIzVdz88PPzu3buWrgJMCcdbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsMTO0gWwQaVS3bt3z9JVAIA1q6mpsXQJbEBu6aWoqMjLy8vSVQAAAM4TAgAAUwQcx1m6hg5NqVSWlZVZugpgzKZNm955552BAwfu3bvX0rUAY6RSqYuLi6Wr6NBwnvAxJBJJr169LF0FMMbJyYkQIhKJ8OYBMDmcJwQAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgiZ2lCwBgWG5ubmFhYfPply5dIoSUlZVlZmY2f9bOzm7y5Mnmrg3AWgk4jrN0DQCs2rx58+zZs9s615gxYw4dOmSOegA6A5wnBDBcXFycWCxu61xJSUnmKAagk0BuARjO2dl53LhxbZrF3t4eJwkBjIHcAjBKWw+eYmJiHB0dzVQMQGeA3AIwyqRJkxwcHPRvj5OEAEZCbgEYxd7e/rnnntOzsYODQ3R0tFnrAbB6yC0AY+l/CBUfHy+RSMxaDIDVQ24BGGvMmDHu7u76tMRJQgDjIbcAjGVnZxcfH//YZu7u7iNGjGiHegCsG3ILwAT0OZCaNm2anR1GqAEwFsbLADABjuP69OnT4phPvNOnTw8ePLi9KgKwWjjeAjABgUAwdepUHQ18fHwGDRrUbvUAWDHkFoBp6D5V+PzzzwsEgnYrBsCK4TwhgMkEBATk5+e3+NSVK1cCAgLauR4Aq4TjLQCTSUxMbHH6008/jdACMBXkFoDJJCcnt3gyELdtAZgQcgvAZORyeWhoqNZEgUDQ2nEYABgAuQVgSs0PrYYMGdK7d2+LFANglZBbAKaUmJhoa2urOQUnCQFMC7kFYEo9evQYPnw4/6ednV1CQoLlygGwQsgtABObNm0a/3jUqFHdu3e3YDEA1ge5BWBiU6ZMEYvF9DFOEgKYHO47NtbFixf1/7eB0EmUlZUplUqBQNCjRw8bG+wdwn+8/PLLS5YssXQVbMPo1Maqr6+/d++epauAjojjuAcPHli6CuhYKioqLF0C85BbJpORkWHpEqCjaGxsnD179ty5czGWLvDWrFlz+vRpS1dhDZBbJoPLxkDTqVOnli9fLpFILF0IdBSZmZmWLsFK4Mw7gFl88MEHCC0Ac0BuAZhF165dLV0CgHVCbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFtgGtu2bQsJCXFwcBAIBAKBYMeOHZauCACsE3ILTGPmzJkXLly4cOGCpQsBACuH3ALocDiO+9e//iWXy999910zraK0tHT9+vVJSUn+/v5ubm4ikUgsFjs5OfXt2zciImLWrFmrVq26cuWKmdZuPo/ddGVlZcOHD3/iiSfs7OwEzYhEom7duvn6+k6ePHn58uX5+fntXD/oA7kF0LEcP3580KBB06ZNKygoMMfya2trFyxY4OXl9fe//93JyWn16tWXLl169OhRRUXFxYsXV61a5evru2fPntdee23jxo3mKMB89Nl0rq6ux48fLywsfOutt+iUTz75hOM4juPq6+vv3Lmzd+/eWbNm3bhxY+nSpQEBAUOGDDl//nx79QD0gtwC6Cjy8/NjYmKio6OdnZ3NtIoHDx6EhoZ+8sknvXr1unLlyoYNG8aPH+/p6SkWi6VSae/evSdPnvzFF18UFhYGBAQ0NjaaqQyTM8mmE4lEHh4ew4YNe/31169du5aenu7g4HDmzJlBgwYxF+HWDbkF0CHk5+enpqZOnjz5zz//3Lx5szlWoVQqJ02adO3aNZlMduTIkX79+rXW0tnZeciQIQ0NDeYow+TMsekEAkFycvKpU6e6du2qUqnmzp37zTffmGTJYDzkFpiSjQ3eUQbq37//kSNHXnrpJUdHRzOt4uOPP/7ll18IIW+88cYTTzyhu3FAQAAruWW+TRcYGPj5558TQtRq9YIFC2pra027fDAMvmXAlIRCoaVLgJY1NDSsXr2aEGJnZ/fSSy89tv2CBQvS09PNX1dHl5iYGBgYSAi5f//+V199ZelygBDkVgdRU1OzcuXKZ5991snJSSQSeXt7T58+PS8vjz6bl5cXHBwsFAoFAkFwcDAh5PTp03PmzPH19ZVIJHK5fP78+QqForWFHzt2bMaMGXK5XCqV2tvbe3h4BAQExMXFZWVl8W1UKtWOHTtiYmLoTx2Ojo6BgYELFy7U8fu2SqXaunVrdHS0h4eHSCRycHCQy+UzZ840rI+EkKysrKCgINrN0NBQQohard69e/f48eOdnZ1ramr03Jim7W9bN/6wYcPEYjG9OE0oFD711FPr1q3jn/3f//1fPz8/ujS5XL5z5049O9Wiqqqq4cOHu7i4bNq06bGNT58+XVJSQggJCAjo3r27wSu1jk2nP4FAkJiYSB//+OOP7bNSeAwOjHP69Gkjt+Tt27d9fX0JIQsWLLh3755CoUhPT5dIJEKhcO/evXyzlJQUQoijo+PQoUObv47x8fHNl9zQ0JCcnEwIGTx48LFjxxQKRWlp6c8//0wXtWrVKtqsvLw8PDycEJKUlHT16tW6uroHDx6kpaVJpVKxWJyent58yWVlZWFhYYSQUaNGnT17VqlUKpXK4uLiffv20Xq2b99uQB/5bsrl8iNHjvTv35/v4KNHjx67Jc3X3zZt/N27d9Pp8+bNa7HOMWPGeHt7NzU1tdaRu3fv0iUsX75cR3/T0tJoM3t7+/r6et0b55133qGNX3zxRd0tdbCaTcdx3LJly2hL/nrC1pw6dYq2dHBw0N1St4SEBELIokWLjFkIcByH3DKW8bk1bNgwQsikSZM0J7799tuEEGdn56qqKjqFfv5lMll+fn5JSUldXZ1CoTh8+LCnpyct4O7du1pLnjNnDiHE39+/trZW66l+/frx3+OxsbGEkKioKLVardmG7tIKhcLc3Fyt2SdMmEAICQkJ0fq6/O2331rMLT37yHdTKBROnDhx7dq1hw8fpjvd+uSW+frbpo2vUqnkcjkhpGvXrtXV1VqL+v333wUCwYcffqijI3p++X700Ue0mVgsfmxuzZo1izZ+4403dLfUwWo2HdeW3Lpz5w4ftEqlUndjHZBbpoLcMpaRuXXixAk6+8GDBzWnX716lU7funUrncLvt2otgb9C99ChQ5rTr127Rq+S4Jegafbs2fR7nP5QTwjZv3+/Vhu1Wk2/RCZMmKA5nd/9zMzM1JqlxdzSv48tdlOlUvn5+T02t8zX3xarolrb+J988klr34mvvfZaly5dKioqdPRFzy/f8vLyiIgIJyenDRs26GhGTZo0iS7z/ffff2zjFlnTpuPakluaJzOb7x3qD7llKvh9y8IOHDhAH9Dffnk+Pj70wdmzZ3UvgZ9RqVRqTs/IyFCr1YSQ0aNHN58rJSVl5MiRhJA9e/bQKUOGDNFqIxAIIiMjCSGHDh3SXHhmZiZ90OKSmzOyjzY2NtevX5fJZLrXYr7+6tDaxp81a5aLiwshZP369RzH8dMbGhq+/PLLF154wcnJSZ/l6+bs7JydnV1RUfHKK688tjF/M5atra1hq7OmTdcmKpWKf2xnZ9fOa4fmkFsWRgeSEQgE7u7umtNlMhn9hJSWlupeAn8Jn+aHnBBC944lEkmPHj2azzVgwICQkBBCyKVLl2izFm/YpDvRTU1NmgPeXLx4kRDi7u6u52XHxvdRH+brrw6tbXypVEqz5Ndffz1y5Ag/fc+ePaWlpQsWLNCvT6bk6upKH1RVVRm2hE676TSvCaKZCpaF3LKwyspKQgjHcfRCKU1NTU2EkPr6esOWTC8e69Kli+5mZWVlOpo5ODhoNtNcMv/UY5mqj35+fs0HlBMIBDSxzNdfw8yfP18kEhFC+BNfhJB//vOf48aNe/LJJ41cuAG8vb3pgwcPHhi2hE676fhzj56enrQwsCzkVocgkUhaO5O7d+9eY5ZcWVlJz54ZpsV5adi09aZU8/VRkzn6axgPD4/nn3+eEPLjjz/Sa8SvX79+4sSJ1NRUU62iTYYPH04f8D9TmZYVbzr+h1h6eRFYHHLLwujJeqVSqf/9SXpyc3MjhDQ1NdEDkdZ069aN/PeZEE3V1dWazSh6erC4uFjzvL8OpurjjRs3Woy9Z555hpizvwZ79dVXCSFqtfqzzz4jhPzzn/986qmnxowZY/ySDRAeHk5v28rPzy8sLHxs+wcPHmgdWnXaTXfw4EH6gL+2BSwLuWVhfn5+9AH98cCE6J2ehBDNHwlaa1ZXV9fi6Z1bt24RQoRCYUBAAD+Rnqupq6s7d+6cPpWYr4+azNdfg/Xv33/cuHGEkM2bN5eWlm7btm3hwoXGL9YwYrH49ddfJ4So1epPP/30se3z8/O1Rj/pnJuuuLh4//79hBAvL6+pU6dapAbQgtyyMP7itxUrVpjwTAshJD4+nj5YsmTJw4cPW2sWFxdHH+Tk5Gg9xXFcdnY2ISQmJkYsFvPTo6Ki6IP3339fn0rM10dN5uuvMehxQ2VlZXx8vI2NzYwZM0yyWKqysjIyMtLFxUXP0cpTUlJGjBhBCFm3bh29uEaH5rllTZtOf4sXL66trbWxsdm8eTMuJuwgkFsWNnLkyIEDBxJCDhw4MHbs2KNHj9JfaKqrq/Pz8/fs2fPo0SPDlhwUFETHpyksLAwNDd24ceMff/xRX1+vUqlKS0vPnTtHx7cODg6ePHkyIWTNmjVaV3bt2LGjsLBQKBQuXbpUc3pCQkKvXr0IIfv27UtKSrp8+XJDQ4NCobh9+/bJkyfbs1LSkEsAACAASURBVI/t019jREVFBQUFEUJOnDjx8ssvSyQSUy2ZELJly5YTJ05UVFSkpqbq83OjUCjcvXv3sGHDGhsbx44de+bMmdZa3rlzJz09vfnxltVsOn00NTW9+uqrdDts2LBBz7s+oD0YdtsX8IwfL+PmzZv8nUxavL29y8vLabPW7t/Mzc2ljb/99lutp6qrq6Ojo1t76WNjY2mz8vJyOmhTYmLi1atX6+vri4qK0tLSJBKJWCzeuXNn85pzcnJavJ6QHwRBa7wMPfuoo5v6MF9/Ddj4vG3bthFC7Ozs9L9lVc+bZ+k4uUS/8TJ4TU1N7777rouLi42NTXJy8nfffXf//v36+vrq6upbt26lp6cnJSXRq+a8vLy05rWaTcdp3He8fPnysrIypVJJBwGprKzMz8//+OOP6fnt0NDQnJwcPdeuG+47NhXklrGMzy2O46qqqt57772hQ4e6uLjY2dm5ublFRUV9+umnNTU1HMddvHgxODiYP0fh6elJh5ijp4n4m6hkMllkZGRDQ4PmktVq9XfffZeQkODj40O/XHx8fOLj43ft2qVSqfhmjY2NX3755ZgxY9zc3Ozs7BwcHAIDA1NTUwsKClqrubCwMCUlpW/fvvb29lKpdMCAAStWrKiurqZ5JpVKw8LCSkpK9Owjx3HHjh0LCgriu+nm5hYWFpaVldWmLWny/hqz8amGhoaePXtOmzZNd+V5eXljx44NCQnx8vKyt7eni7WxsXFzc+vfv//w4cPXrl2rNUtFRYX+42Voqamp+eqrr1544YX+/fvTjSASiZydnf39/ceOHbt48eKvv/76wYMHzWdkfdOVl5fHxMQEBAQ4OjrSETLt7OwEAgG/fyOTyby9vcPDw994440TJ060dcPqgNwyFQH338f70FZnzpyhwwdgSwKADlOnTs3MzFy0aBE/IDIYBr9vAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMASO0sXYD28vLwsXQIAdFxlZWWWLsFKILdM5t69e5YuAQDA+iG3jOXr65uRkWHpKqBj2b9//9atW319fd99911L1wIdi6+vr6VLYB5yy1guLi4JCQmWrgI6Fnrw7erqivcGgMnhugwAAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAlthZugAAhikUioaGhubTlUolIaSpqamioqLFGZ2dnc1bGYD1EnAcZ+kaAFj1wQcfvP76622dKyQk5Pz58+aoB6AzwHlCAMMlJycLBAID5jJHMQCdBHILwHBeXl5hYWFtmsXGxiYxMdFM9QB0BsgtAKMkJSW1qX1ERESvXr3MVAxAZ4DcAjBKQkKCUCjUv31bcw4AtCC3AIzi5uYWFRWlZ2OhUBgfH2/WegCsHnILwFj6H0KNGzfO1dXVrMUAWD3kFoCxYmNjpVKpPi1xkhDAeMgtAGN17dp1woQJj20mlUonTpzYDvUAWDfkFoAJ6HMgFRsbK5PJ2qEYAOuG3AIwgQkTJri4uOhug5OEACaB3AIwAZFIFBsbq6OBs7PzmDFj2q0eACuG3AIwDd2HUwkJCSKRqN2KAbBiyC0A0xg5cqSnp2drz+IkIYCpILcATMPGxiYhIaHFpzw9PcPDw9u5HgBrhdwCMJnWDqqSkpJsbW3buRgAa4X/vwVgSr6+vr/99pvWxF9++WXgwIEWqQfA+uB4C8CUpk6dqjWlb9++CC0AE0JuAZjS888//9gpAGAMnCcEMLEBAwZcunSJ//PGjRtPPvmkBesBsDI43gIwMc2rMwYOHIjQAjAt5BaAiSUlJQkEAv6xZYsBsD52li7ANM6cOZOWlmbpKgD+zdXVtbS0VCAQHD9+/OzZs5YuB4AQQoYOHZqammrpKkyBswoZGRmW3pAAAB1aQkKCpb+qTcNKjrcomUz217/+1dJVABClUvn555+PHDkyMDDQ0rUAkOzs7AsXLli6CpOxqtxycnLC2ULoIB48eLBlyxZXV1dLFwJAFi9ebE25hesyAMxi3bp1CC0Ac0BuAZjFE088YekSAKwTcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILesyrZt20JCQhwcHAQCgUAg2LFjh6UrsrysrKzBgwd369aNbpPU1FRLVwRthhcRNCG3rMrMmTMvXLhw4cIFSxfSgYwYMeLMmTOlpaVdunSxdC1gILyIoAm5BfBvHMf961//ksvl7777rkkWWFZWNnz48CeeeMLOzk7wf7Kzs/WZd9OmTfwsjo6OTz/99MqVK01SVVsLFggENjY2jo6Offr0mTJlypYtW2pra81dSVvdvHnT39/f09Pzp59+smAZj30LtbaFKZFI1K1bN19f38mTJy9fvjw/P7+d62cFcguAEEKOHz8+aNCgadOmFRQUmGqZrq6ux48fLywsfOutt/iJ69ate+yMarX6ww8/5P+8f//+5cuXX3/9dVMV1prmBX/yySccxykUiry8vGXLll29evWll17y8/P7+eefzV1Mm+zatevGjRtFRUUbNmywVA36vIVa28Icx9XX19+5c2fv3r2zZs26cePG0qVLAwIChgwZcv78+fbqATOQW9DZ5efnx8TEREdHOzs7m3VFIpGIEPL9998XFhbqbpmZmXnr1i3aviOQSCS9e/d+4YUXcnNz/fz87t69O378eBMGvPGmTp3ar18/d3f3v/3tb+2/dpO8hUQikYeHx7Bhw15//fVr166lp6c7ODicOXNm0KBBGzduNGG1VgC5BZ1afn5+amrq5MmT//zzz82bN5t1XQkJCYQQlUq1fv163S1Xrlw5YsQILy8vs9ZjgK5du/6///f/CCFVVVUffPCBpcv5Dz8/v5s3bz58+HDMmDHtvGpzvIUEAkFycvKpU6e6du2qUqnmzp37zTffmGTJ1gG5ZYVsbPCy6qt///5Hjhx56aWXHB0dzb2ulJQU+tJs3rxZoVC01uzAgQN5eXn/+7//a+56DBMREUEfWPaXJAu6d++e5p/mewsFBgZ+/vnnhBC1Wr1gwYIO+LOipeALzgoJhUJLlwAt6N27d0xMDCGksrLyq6++aq3Z+++/P2DAgLFjx7ZfZW3h7u5OH2h9fXcG+fn5iYmJQUFBpaWl7bPGxMTEwMBAQsj9+/d1vGc6G+TWf9TU1KxcufLZZ591cnISiUTe3t7Tp0/Py8ujz+bl5QUHBwuFQoFAEBwcTAg5ffr0nDlzfH19JRKJXC6fP3++jp3oY8eOzZgxQy6XS6VSe3t7Dw+PgICAuLi4rKwsvo1KpdqxY0dMTIynp6dYLHZ0dAwMDFy4cKGOHxJUKtXWrVujo6M9PDxEIpGDg4NcLp85c6ZhfSSEZGVlBQUF0W6GhoYSQtRq9e7du8ePH+/s7FxTU6NjA547dy4wMNDW1lYgEMTFxR05ciQhIaFnz54ikcjd3f25555rbQ/dsI63dRZTqaqqGj58uIuLy6ZNm9o674IFC+iDjz/+mOO45g1ycnJOnjypz8HW9evXp0yZEhQU1KNHD4lEIpPJ+vXr9/LLLz948IA2KCgoCA0NFYvF9Fo1sVj8zDPPXLt2jRDy9ttvy+VyGxsboVAYHBx8+PBh/btQVVVFH4jFYq2nLPU6xsfHe3p60m7OmzePTjTmA9vQ0LB+/frIyEhXV1ehUOjk5PTMM8+MGzcuMDDwwIEDP/74Y7du3fQvzxgCgSAxMZE+/vHHH9tnpQzgrEJGRgYhpFevXgYv4fbt276+voSQBQsW3Lt3T6FQpKenSyQSoVC4d+9evllKSgohxNHRcejQoc03Znx8fPMlNzQ0JCcnE0IGDx587NgxhUJRWlr6888/00WtWrWKNisvLw8PDyeEJCUlXb16ta6u7sGDB2lpaVKpVCwWp6enN19yWVlZWFgYIWTUqFFnz55VKpVKpbK4uHjfvn20nu3btxvQR76bcrn8yJEj/fv35zv46NGjx27JCRMmEEIkEkmL77d//OMfWu0N6LgBs3AcR2/9WbhwYWuV3717lxa5fPlyHR1MS0ujzezt7evr6x+7QTiOW7ZsGSGkqKiI4zh+e/7444/NW44fP75Pnz5NTU0cx8nlch2b/cCBAzY2Nh999NG1a9cUCsWtW7eioqIIIT4+PlVVVXyz8+fP0+PvuLg4zdkvX74slUp37dqlo2CicbUb79y5c/Sp0NBQzent8zrqeBFdXV0JISkpKZoTDfjAVlRUhISEEELGjRt38eJFpVL5+++/T5o0ic6ydevWFrcYpedbiNO5hbWcOnWKtnRwcNDdUodFixYRQhISEgxeQoeC3Pq3YcOGEUImTZqkOfHtt98mhDg7O/NfBPRjIJPJ8vPzS0pK6urqFArF4cOHPT096Xvr7t27WkueM2cOIcTf37+2tlbrqX79+vG5FRsbSwiJiopSq9WabXbu3EkIEQqFubm5WrPThAgJCdH69vztt99azC09+8h3UygUTpw4ce3atYcPH6ZXb+ufWz179ty7d+/9+/fr6+uLi4s3btzo4OBAqzp06JBmewM6bsAsnOly66OPPqLNxGKxAbnFX6g9duxYrWb0wPezzz6jfz42t0aMGKE55ddff6XtN23apDn9nXfeodP5JKisrOzbt++iRYt0F9zityof2++9957m9PZ5HQ3LrTZ9YP/6178SQnr37t3Q0MBPbGpqoi9HdHR081XzzJFbd+7c4YNWqVTqbtwa5FZHZGRunThxgr4tDh48qDn96tWrWjtZ/O6b1hL4C1W1vpSvXbtGf4pvcTdt9uzZNLd++eUXOvv+/fu12qjVavqBmTBhguZ0fi8sMzNTa5YWc0v/PrbYTZVK5efnp39uDRw4UGv6wYMH6YpGjhzJTzSg4wbMQpkqt8rLyyMiIpycnDZs2KCjmSbN3KqpqXFyciKECASC69evazabNm2au7s7/92kO7eaU6vV9NDqtdde05ze1NT07LPPEkKcnZ0fPHigVqsnTZoUERHR2Niou+Dm36r37t3r3r07IUQul2vu6LTb62jw8ZZW49Y+sPX19XQVc+fO1ZqF7oCKRCKVStV87ZQ5ckvzZGbzlNWTleUWft8ihJADBw7QB/QnUJ6Pjw99cPbsWd1L4GdUKpWa0zMyMtRqNSFk9OjRzedKSUkZOXIkIWTPnj10ypAhQ7TaCASCyMhIQsihQ4c0F56ZmUkftLjk5ozso42NzfXr12UymT7ratHYsWODgoIIIdnZ2fyVUQZ03IBZTMvZ2Tk7O7uiouKVV14xYPYuXbrMmjWLEMJx3Mcff8xP//333zMzMxcuXGhvb29YYXTABUJIfX295nRbW9vt27dLJJKKiorZs2d/8MEHubm5//rXv+zs7PRccl1d3e+//75+/frQ0NCHDx8GBQUdPnyYP4AmDL6OrX1g6flzQkiPHj20ZqGHaA0NDfwviO1DpVLxj/V/yawbcosQQuh4KgKBgL9WipLJZPSN8tjLh/hL+Lj//rGd7lRKJJLmHwNCyIABA+iZ9EuXLtFmLd63SPc9m5qaNMd9uXjxIiHE3d1dz6tvje+j8eiJSpVKRa8OIAZ13IBZOpp58+bRo/Bt27ZVVFTQiR9++KFUKp07d6451ujr60tvt/rxxx+XLVuWmZnp4eGhz4zz588XCAQSiSQ4ODgtLS0iImLXrl3nz5/v06ePZjPmXsfWPrC2trb0QV1dndYsdAeUVmWOklqjeTGUi4tLe666w0JuEUJIZWUlIYTjOHr1kSb6HtXagdVfSUkJIeSxg4GWlZXpaMbv2NJmmkvW3OfVzVR99PPzaz6umkAg4E/76NCzZ0/6gP+yNqDjBszS0fTu3ZueTVUoFF988QUh5MGDB1u3bv3b3/5GTyHqo6qqas2aNdHR0X379nV3d5fJZEKhUMcVrfPmzRs1ahQhRCgU6hlaROMsVnV1dUFBwa5duxITE/kvd57VvI49evSgx7tFRUVaT9HfmWQyGf82bh/8uUdPT8+OM4SKZSG3/kMikbR2OnXv3r3GLLmyspLfWTNAi/PSsGloaGjToszXR33wO7l6nu4wYKMZs53bE39B/Pr161UqVVpaGsdx+v97jvv374eEhCxevFgqlW7ZsuXixYslJSVKpVLHHlJjYyPdc1coFC+++GJ7biiGXkeRSDR58mRCyL59+zT35BobGw8dOkQImTZtWjvfH8n/Ak1PVwBBblF0J1epVOq+P8kAbm5uhJCmpiZ6eNQaejtIa2uvrq7WbEbR04PFxcWap791MFUfb9y40WLsPfPMM4+d99GjR/QBf67SgI4bMEsHFBUV5e/vTwi5c+fOli1bNm7cOGPGDP135JctW1ZQUBAQEJCRkREREdGzZ0+JRKJ7b2DRokWPHj365JNPCCEnTpxYu3at8b3gWdPrmJaW5u3tXVJSMnny5CtXrigUiitXrkyaNOnevXsDBgzQHO+4ffAXNPHX4gNyixBC/Pz86AN6zt2E6A2PhJAjR448tlldXV2LZ0Vu3bpFCBEKhQEBAfzEJ598ks7C30+jm/n6qD/akS5duvDFGNBxA2bpmObPn08fpKSk1NbW/s///I/+89KLaIKDg5ufsmvRjh07tm/f/s0338ybN4+eonzzzTdv3LjR9qpbZk2vo4eHx/HjxyUSSUFBwdChQx0dHSMiIqqrq9esWXPmzBlzD76spbi4eP/+/YQQLy+vqVOntueqOzLkFiEal+StWLHCtCco4uPj6YMlS5Y8fPiwtWZxcXH0QU5OjtZTHMfR/9gUExOjOUIBvcmUEPL+++/rU4n5+qgnlUpFBweJi4vjv20N6LgBs5hWZWVlZGSki4uLkaN0z5w5kx40NzY2xsbG0h0RPdGrBv788099Gl+5cuWVV1758ssv6So+//xzJyenurq6mTNn6nmw/lgsvo6tefjwYWxs7ODBg2/cuPHo0aOmpqaKiopTp06lpqa2fzGLFy+ura21sbHZvHkzLibkIbcIIWTkyJEDBw4khBw4cGDs2LFHjx6lv0hVV1fn5+fv2bOHP8HVVkFBQXSYlsLCwtDQ0I0bN/7xxx/19fUqlaq0tPTcuXN0mOfg4GB6Vn3NmjVaFzjt2LGjsLBQKBQuXbpUc3pCQkKvXr0IIfv27UtKSrp8+XJDQ4NCobh9+/bJkyfbs4962rp1a1FRkUQiWbJkCT/RgI4bMItpbdmy5cSJExUVFampqW39fVETf0E8IaSto+jSy/mOHj26fft2hUJRU1Nz8uTJ2bNnN79qvKqqKj4+fs6cOfwulKenJ72LPDc3V8+dnsdi8XVs0aNHj8LDwy9fvkyPsdp57ZqamppeffXV9PR0oVC4YcMGPW936SzaeL9XB2X8eBk3b97k72TS4u3tXV5eTpu1dhtjbm4ubfztt99qPVVdXR0dHd3a9o+NjaXNysvL6aBNiYmJV69era+vLyoqSktLk0gkYrF4586dzWvOyclp8XpCfiwArfEy9Oyjjm7qg56Gov9Xorq6uqmp6eHDh+vWrZNIJFKp9IcfftBqb0DHDZiFM919x6tXr6bNDBsvQ9OtW7dsbGy0hr3g6bjv+Ntvv23+Ig4ePJjmmWYfY2Njhw4dSgeO0kSP14VC4S+//NJawUSPu2J57fM6muq+49Y+sE1NTXQstOZsbW3d3d0nTZp05syZ1jaCAfcdL1++vKysTKlU0kFDKisr8/PzP/74Y3ouPTQ0NCcnR/ei9GFl9x0jt/6jqqrqvffeGzp0qIuLi52dnZubW1RU1KefflpTU8Nx3MWLF4ODg/lDdU9PzxdffJHjOHrWiL+JSiaTRUZGao4Qw3GcWq3+7rvvEhISfHx86GfSx8cnPj5+165dmvfeNzY2fvnll2PGjHFzc7Ozs3NwcAgMDExNTS0oKGit5sLCwpSUlL59+9rb20ul0gEDBqxYsaK6uprmmVQqDQsLKykp0bOPHMcdO3YsKCiI76abm1tYWFhWVpb+m5Hmlru7+8CBA11cXIRCoVQq9ff3T01NvXPnTouzGNBx/WfJysoaOnQo//O+ra2tXC7nR+vJy8sbO3ZsSEiIl5cXf8OvjY2Nm5tb//79hw8fvnbtWq0FVlRU6DleRllZ2ahRo3r37k23Z9euXQMDA+fNm6fZJiYmRmsEk40bN4aGhmpdgxMcHPzhhx9qNfPz86P/2T0iIuKLL75oamqiP4EIhUJ6ayD/Y0yvXr327dvHbxB+z4aSy+V//vknfXbatGl9+/bl3wB2dnZeXl5DhgyhVzzqZr7XUceLmJOTExYWRi+Aok/16dNn3Lhx58+fN+wD++eff/JD2bZIJBJpZkmb3kLl5eUxMTEBAQGOjo50aFA7OzuBQMAvXCaTeXt7h4eHv/HGGydOnHjsNtcTcqsjMklugUnQ3IqIiLB0IQCG+O233+Ry+dKlS7Wm19fXHz16lN75q3uUwg7IynILv2+BWZhvmCUA86msrBw1alR9ff2bb76p9ZRIJBo5ciT9SZIf8AUsArkFZsHK/b8AmtavX3/nzp3nnnuutZEp6BCd+o9TA+aA3AIA+Df6v2B0DAN4/Phxovdg1mAmyC0AgH+jd0O3eCcJIWTLli3Hjx/v3r37a6+91r51wX9BbgEA/NvcuXMjIyOzs7NnzZp1/vz52tpaOkjbwYMHp06d+tJLLwUHB9PosnSlnRpyC0zm66+/9vPzo//o6/z5856enuHh4fR/+AIwwd7ent7NXVpaGhcX5+rqKpFIfH19//73v0skku+++y43N5cfpQwsBQOHgMkkJycnJydbugoAo9ja2k6fPn369OmWLgRaheMtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWGJn6QJMqaysbOrUqZauAgCgY7l06ZKlSzAlq8otpVKZmZlp6SoAAMCMrCS3fH19Fy1aZOkqAP7twoUL2dnZPXr0mDZtmqVrAfi3AQMGWLoE0xBwHGfpGgCszZo1axYvXjxkyJCff/7Z0rUAWBtclwEAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwRMBxnKVrAGDV6tWrP/vss+bTq6qqysrK7O3tPT09mz/7xBNPHD161PzVAVgnO0sXAMCwkSNH/v3vf2/t2bq6uoKCgubTp06das6iAKwcjrcAjOLv73/jxo02zXLp0qWnn37aTPUAWD38vgVglKSkpDa19/f3R2gBGAO5BWCU559/XiAQ6N9++vTp5isGoDPAeUIAYz377LO5ubn6tBQIBL///nvv3r3NXRKAFcPxFoCx9D9VOHjwYIQWgJGQWwDGmjZtmq2trT4t2/pjGAA0h9wCMFaPHj0iIiIe28zW1hZXwAMYD7kFYAL6HEiNGjWqe/fu7VAMgHVDbgGYwJQpU8Rise42OEkIYBLILQATcHZ2Hjt2rI4G9vb2sbGx7VYPgBVDbgGYhu7DqQkTJjg5ObVbMQBWDLkFYBrPPfecTCZr7VmcJAQwFeQWgGlIpdJJkya1+JSDg8P48ePbuR4Aa4XcAjCZ1g6qJk+eLJFI2rkYAGuF3AIwmTFjxnTr1q35dJwkBDAh5BaAyQiFwilTpmhNdHNzGzVqlEXqAbBKyC0AU2p+aJWYmGhnh3/QCmAyyC0AUwoPD/fx8dGcgpOEAKaF3AIwJYFAkJCQwP/p4+MzZMgQC9YDYH2QWwAmpnmAlZyc3Kb/KgkAj4XT7p3amjVrLF2CderevfvDhw8JIU1NTdjI5hATE9OvXz9LVwGWgf933KnhUAAYlZGRgzLdXgAAIABJREFUoXk+FjoVHG8BCQgI0DFAERigrq4uLy/Py8urZ8+elq7F2ly4cKGhocHSVYAlIbeAbNq0afDgwZauwto8++yzX3/9dd++fS1diLXx8vK6d++epasAS8J1GQBmsXr1aoQWgDkgtwDMIjw83NIlAFgn5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW4BAABLkFsAAMAS5BYAALAEuQUAACxBbgEAAEuQWwAAwBLkFgAAsAS5BQAALEFuAQAAS5BbAADAEuQWAACwBLkFAAAsQW5BZ7dt27aQkBAHBweBQCAQCHbs2GHpigBAF+QWdHYzZ868cOHChQsXLF0IAOjFztIFAIC2srKy+Pj4wsLCe/fuqVQqrWeFQqGDg4OLi0v//v1DQkLi4uL69+9vkToBLALHWwAdjqur6/HjxwsLC9966y065ZNPPuE4juO4+vr6O3fu7N27d9asWTdu3Fi6dGlAQMCQIUPOnz9v2ZoB2g1yC4AlIpHIw8Nj2LBhr7/++rVr19LT0x0cHM6cOTNo0KCNGzdaujqA9oDcAmCVQCBITk4+depU165dVSrV3Llzv/nmG0sXBWB2yC0AQgixsWH1sxAYGPj5558TQtRq9YIFC2pray1dEYB5sfpZBTAtoVBo6RIMl5iYGBgYSAi5f//+V199ZelyAMwLuQVGqampWbly5bPPPuvk5CQSiby9vadPn56Xl0efzcvLCw4OFgqFAoEgODiYEHL69Ok5c+b4+vpKJBK5XD5//nyFQtHawo8dOzZjxgy5XC6VSu3t7T08PAICAuLi4rKysvg2KpVqx44dMTExnp6eYrHY0dExMDBw4cKFBQUFrS1WpVJt3bo1Ojraw8NDJBI5ODjI5fKZM2ca3M2srKygoCDazdDQUEKIWq3evXv3+PHjnZ2da2pqCCFVVVXDhw93cXHZtGlTG7avfgQCQWJiIn38448/tql4g18jhUKxatWqgQMHuri4dOnSpXfv3lFRUUuXLm3T2gEMwUEnRt8Dp0+fNmz227dv+/r6EkIWLFhw7949hUKRnp4ukUiEQuHevXv5ZikpKYQQR0fHoUOHNn8HxsfHN19yQ0NDcnIyIWTw4MHHjh1TKBSlpaU///wzXdSqVatos/Ly8vDwcEJIUlLS1atX6+rqHjx4kJaWJpVKxWJxenp68yWXlZWFhYURQkaNGnX27FmlUqlUKouLi/ft20fr2b59uzHdlMvlR44c0bww/dGjRxzHpaWl0T/t7e3r6+v13MLLli2jc/HXE7bm1KlTtKWDg4PBxev/GlVVVfn7+wsEgnXr1pWUlJSWlh46dCg4ONjV1dWAtbdJr169CCEZGRmGzQ5WALnVqRmZW8OGDSOETJo0SXPi22+/TQhxdnauqqqiU+h3okwmy8/PLykpqaurUygUhw8f9vT0pAXcvXtXa8lz5swhhPj7+9fW1mo91a9fPz63YmNjCSFRUVFqtVqzzc6dOwkhQqEwNzdXa/YJEyYQQkJCQrTC47fffmstt9rUTaFQOHHixLVr1x4+fHjdunV8bn300Ud0+WKx2By5defOHT5mlEqlYcXr/xq98847hJDJkydrTjx+/LhWbum59jZBbgFyq1MzJrdOnDhBZz948KDm9KtXr9LpW7dupVP4fXmtJfDXbR86dEhz+rVr1+hVEvwSNM2ePZvm1i+//EJn379/v1YbtVotl8sJIRMmTNCczh+UZGZmas3SWm4Z002VSuXn50dzq7y8PCIiwsnJacOGDc071Rr9c0vzVB6fMeZ7jWJiYgghKSkpmhMbGxv79evH/6n/2tsEuQX4fQsMdODAAfqAXhHA8/HxoQ/Onj2rewn8jEqlUnN6RkaGWq0mhIwePbr5XCkpKSNHjiSE7Nmzh04ZMmSIVhuBQBAZGUkIOXTokObCMzMz6YMWl9wiY7ppY2Nz/fp1mUxGCHF2ds7Ozq6oqHjllVf0XHWbaA6rYWf373FwzPcadenShRCSkZFx5coVzfXevHmT/9P4tQO0CLkFBsrPzyeECAQCd3d3zekymYx+b5aWlupeAn8JH/d/R34UPZCSSCQ9evRoPteAAQNCQkIIIZcuXaLNnJ2dmzejx1tNTU20TurixYuEEHd3d0dHx8f179+M72b7oFd/UC4uLvSB+V6j559/nhBSUlISGhr6l7/85aeffmpqatKal5VNB8xBboGBKisrCSEcx9FL0TTRr7D6+nrDllxSUkL+b49eh7KyMh3NHBwcNJtpLpl/Sh/m66Zp3b17lz7w9PQUiUT0sfmKnzhxYlpaGr3GZOvWraNHj+7Zs+cbb7zx6NEjvg0rmw6Yg3F1wSgSicRMN7pWVlaq1WqDbwemZxq10K/LhoaGti7NfN00Ff5HI3ophCYzFb9o0aIpU6Zs27YtMzPz0qVLxcXFK1eu3LVr17lz59zc3My9dujMcLwFBnJyciKEKJVKzTNUJkG/9ZqamujhUWu6detG/vv8mKbq6mrNZhQ9PVhcXNx8kPXWmK+bpnXw4EH6YNKkSfxEcxfv5eX15ptv5uXl/frrr9OmTSOEFBYWLl68uH3WDp0WcgsM5OfnRx/Q35lMiN79Sgg5cuTIY5vV1dVpngnk3bp1ixAiFAoDAgL4iU8++SSd5dy5c3oWY75umlBxcfH+/fsJIV5eXlOnTuWnt1vxvr6+O3funDx5MiGEvyuciU0HLEJugYH4S/JWrFjR4kk5g8XHx9MHS5YsefjwYWvN4uLi6IOcnBytpziOy87OJoTExMSIxWJ+elRUFH3w/vvv61mMqbpZWVkZGRnp4uJijlHbFy9eXFtba2Njs3nzZv5iQmLO12jixImaF7xQI0aMIIQIBAJzrx06OeQWGGjkyJEDBw4khBw4cGDs2LFHjx6lv0hVV1fn5+fv2bNH8yf6NgkKCqKjFhUWFoaGhm7cuPGPP/6or69XqVSlpaXnzp2jo54HBwfTHfw1a9ZoXe22Y8eOwsJCoVCoNexQQkICvftn3759SUlJly9fbmhoUCgUt2/fPnnypFm7uWXLlhMnTlRUVKSmphrwA1trmpqaXn311fT0dKFQuGHDBq3r+833GnEc13wgRHpNPB3BxKxrh87OYneOQQdA3wMGj5dx8+ZN/l4cLd7e3uXl5bRZa/e05ubm0sbffvut1lPV1dXR0dGtvWljY2Nps/LycjpoU2Ji4tWrV+vr64uKitLS0iQSiVgs3rlzZ/Oac3JyWryekB8Yovl4GUZ2k1q9ejWdxbDxMpYvX15WVqZUKunIIJWVlfn5+R9//DE9FxcaGpqTk9PiEsz0Gk2YMEEoFL755pu//fZbfX39/fv3P/roI1tbW3d3999//72ta28T3HcMyK1OjX6DGJxbHMdVVVW99957Q4cOdXFxsbOzc3Nzi4qK+vTTT2tqajiOu3jxYnBwMH/mytPT88UXX+Q4jp4042+ikslkkZGRDQ0NmktWq9XfffddQkKCj48PzSEfH5/4+Phdu3apVCq+WWNj45dffjlmzBg3Nzc7OzsHB4fAwMDU1NSCgoLWai4sLExJSenbt6+9vb1UKh0wYMCKFSuqq6tpnkml0rCwsJKSEv27eezYsaCgIL6bbm5uYWFhWVlZmkuoqKjQf7yM8vLymJiYgIAAR0dHOpqfnZ0df/6NbjFvb+/w8PA33njjxIkT7f8aff/999OnT3/66aednJxsbW27dOnSv3//V199taioqE1rNwByCwTcf59ggU6FfhWePn168ODBlq4FQC9eXl737t3LyMhISEiwdC1gGfh9CwAAWILcAgAAliC3AACAJcgtAABgCXILAABYgtwCAACWILcAAIAlyC0AAGAJcgsAAFiC3AIAAJYgtwAAgCXILQAAYAlyCwAAWILcAgAAliC3AACAJcgtAABgCXLr/7d373FRlvn/x6+BGWAQOSmomLCGGCSoYJZHNCXN1FUhRNFqt3UrYz1/t9btYFtr+fhqqJ3MzqamX83NtjVPeT7godJM1NQ1PCsnAUEOAvP74/ru/OY7wDgMc+CC1/MPH8M9133P576umfs99z33fQsAUAm5BQBQCbkFAFAJuQUAUAm5BQBQCbkFAFAJuQUAUAm5BQBQCbkFAFAJuQUAUAm5BQBQCbkFAFAJuQUAUInW1QXA9VavXp2RkeHqKgCrFBcXu7oEuBi5BbF48WJXlwAA1iK3mrVevXq5uoSm6erVq+fPn2/ZsmWXLl1cXUvT1KpVK1eXAJfRGAwGV9cANDULFy6cOXNm79699+/f7+pagKaG8zIAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKiG3AAAqIbcAACohtwAAKtG6ugBAYZ999tm3335bc/rp06flv2PHjq35bEBAwNKlSx1eHNBEkVuA7dq3b7927dq6ns3Ly6v12UmTJjmyKKCJ4zghYLtBgwa1adOmvnONHz/eEcUAzQS5BdjO3d291iOBFrRr127AgAEOqgdoDsgtoEHqu/M0fvx4d3d3BxUDNAcag8Hg6hoAtUVERJw9e9bKxocPH77vvvscWg/QtLG/BTSU9YcKw8PDCS2ggcgtoKEmTpxo95YA6kJuAQ0VFRUVExNjTcv6nsQBoCZyC7ADa87OiIuLu/fee51QDNC0kVuAHUyYMEGj0Vhuw2VbgF2QW4AdhIaG9u7d20IDjUaTnJzstHqAJozcAuzD8u5UfHx8WFiY04oBmjByC7CPlJQUrbbOG35ykBCwF3ILsI+goKDBgwfX+pROp0tKSnJyPUBTRW4BdlPXTtWQIUNat27t5GKAporcAuxmzJgxer2+5nQOEgJ2RG4BduPr6zt8+HCzid7e3qNGjXJJPUCTRG4B9lRz1+q3v/2tj4+PS4oBmiRyC7CnRx55xN/f33QKBwkB+yK3AHvy8vIaPXq08c+AgICHH37YhfUATQ+5BdiZ6Q7Wo48+6uHh4cJigKanzssk7ejMmTN5eXlOeCGgMfDx8QkMDMzPzxdCdOvW7cCBA66uCHCeXr16Ofw1DI7HbdkAoJlwQqZwnBAAoBJnHCeUZsyYkZ6e7rSXA1wrIiJi1apV9913n6sLAZzhwIEDlv9LBDtifwtwiNmzZ/fo0cPVVQBNkPP2t4Bm5cknn3R1CUDTxP4WAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAlSubW559/HhcX5+vrq9FoNBrNihUrXF3RnalYMwA0Qkrm1uOPP/7jjz/++OOPri6kHlSsGQAaISVzC83NokWLNPX3zjvv1Lo0g8HwP//zP+Hh4X//+98dVHBubu4777wzfvz4qKiooKAgDw8PT09Pf3//Tp06xcfHP/nkk/Pnz//5558d9OqOY7nr8vLyBg4c+Jvf/Ear1dYcDg8Pj9atW3fu3HnMmDGvvfZaZmZmXa9i3+FG00NuQQ1eXl7btm07ffp0Tk5OWVlZZWWlwWCQT7Vp06a0tPTmzZt5eXmXLl06duzYpk2b3N3da13Ozp07H3jggXHjxp07d84Rdd66dWvq1KkdOnT4r//6L39//zfffPOnn366efPmjRs3jhw5Mn/+/M6dO69bt+65555bunSpIwpwnDt2XatWrXbu3JmVlfXiiy/KKW+//bbBYDAYDOXl5RcuXFi/fv2TTz556tSpl19+OTo6unfv3j/88EOti7LXcKNJ0rq6AMAqHTp0GDRoUF3Penl5GR+3b98+JiamXbt2Zm0yMzOff/75bdu2xcfHO6jIK1euPPTQQydOnOjUqdO3334bERFh+mzHjh07duw4ZsyY+fPnx8fH375920Fl2F3Du87Dw6Nt27Zt27bt16/f888/v2rVqsmTJx84cOCBBx549913n376abP2DR9uNGHsb0ENd911V0PaZ2ZmTp8+fcyYMdeuXfv444/tWtr/Ki0tHTVq1IkTJ3x8fLZu3WoWWqYCAgJ69+5dUVHhiDLszu5dp9FoUlNT9+7d27Jly6qqqmefffYf//iHWZsGDjeaNoVzy81NveJVrLmRqO+GqUOHDqZ/dunSZevWrX/4wx/8/PzsWtf/99Zbb33//fdCiNmzZ//mN7+x3Dg6OlqV3HJQ18XExHzwwQdCiOrq6qlTp966dcv02QYON5o2hY8T6nQ6V5dQbyrW3BhMnz69vrOsWbPGEZXUpaKi4s033xRCaLXaP/zhD3dsP3XqVMcX1dilpKS8/vrrP//88+XLlz/77LNnn31WTm/8ww3XUuPrf1VV1bJly4YNG9a2bVsPDw9fX9/w8PDHH3/cwizFxcXz5s27//77/f39PTw8QkNDJ06cePTo0Zott2/f/thjj4WHh3t7e3t5ebVt2zY6OjoxMXHHjh1mNaxYsWLEiBEhISGenp5+fn4xMTHTpk2r6zdqu9e8Y8eObt266XQ6jUbTs2dPIUR1dfWXX375yCOPBAQEFBcXW+5DB63v0aNHY2NjZVWxsbFCiIyMjMmTJ3fu3Fmv14eHh0+ZMqWkpMR0ln79+nl6espzwHQ63b333rt48WLjs88//3xkZKRcYHh4+KpVq6xcL/sqLCwcOHBgYGDghx9+aE37jIyMnJwcIUR0dHSbNm1sft1m1e0ajSYlJUU+/vbbbxu4NGs+ICdPnnz00Ue7devWrl07vV7v4+MTERHx1FNPXblyRS7Eho6VSkpK5s+f36NHj8DAwBYtWnTs2DEhIeHll1+Wzx46dCgmJsbd3V2j0SQmJm7dujU5Obl9+/YeHh7BwcG//e1vv/vuu1pXqr6bHctlGFm/eWykDI6XnJwshJgxY4Zts+fl5fXt21cIMXjw4IMHD5aWlpaWlmZnZ3/zzTdyFZYvX242y/nz5zt37iyEmDp16qVLl0pKSlauXKnX63U63fr1643NKioqUlNThRC9evXavn17SUlJbm7u/v3709LShBDz5883tszPz+/fv78QYvz48cePHy8rK7ty5Up6erq3t7enp+fKlSudVrOsLTw8fOvWrV26dDGO482bN+/Yk45bX7kEPz+/Pn361HyPJSUlmbX/8ssv5VN/+tOfai11yJAhoaGh8iwyC+RC2rRpc8d1N3Xx4kU542uvvVZXm/T0dNnGy8urvLz8jst89dVXZfvf//739SrGVOPvdmu6zmAwzJkzRzYznk9Yl71798qWvr6+llsarBtuyx+QjRs3urm5LViw4MSJEyUlJWfPnk1ISBBChIWFFRYWmi3E+o4tLCyMiorSaDSLFy/OycnJzc3dvHlzbGxsq1atTJsNHz5cCKHX62suUwjxyiuvmC22vu8HK8uwclNTXxkZGXJFbF6C9RTILTnYcXFxZpuPM2fOyG6qmQH9+vUTQowaNcp04t/+9jchREBAgPENOnnyZCFEVFTUrVu3zJYQERFhuh0fPXq0ECIhIaG6utq0mfxaqtPpDh8+7Jya5SdKp9ONHDly0aJFW7Zskd+arcktx62vrMrHxyczM1OeuFxSUrJly5aQkBC5vhcvXjRtX1VVFR4eLoRo2bJlUVGRWTH//ve/NRrNf//3f99xjazZkNVkzcZ3wYIFso2np6c1ufXkk0/K9rNnz65XMaYaf7fbPbcuXLhg3HCXlpZabmx9btX1Adm4ceODDz5o2v6XX36Ri/3www/NFmJ9x8pvLWPGjDGduHPnzlpzq3379uvXr798+XJ5eXl2dvbSpUt9fX3lYjdv3mzavr7vByvLsHJTU1/k1v9n/Dq2du1as6fqyoDdu3fL6Zs2bTKdfvz4cTl92bJlBoPhxIkT8iwJ+aeZSZMmGbfj8sd2IcSGDRvMmlVXV8sNwfDhw51Qs8Hkm6CxTVVVVWRk5B1zy3HrW2tVkvESJbMPpMFgePvtt+vatD333HMtWrS4ceOG5TUyODK38vPz4+Pj/f39lyxZYs0yR40aJZf5xhtv1KsYIyW63e65ZXrMzSwMarI+t6z/gFRXV8tfnZ977jkLC5Hq6tgRI0YIIdLS0kwn3r59OyIiwnSKzK0ePXqYLXbTpk1ysYMGDTJOtOH9YE0Z1m9q6suZudXYf99au3atfPDQQw9ZOcvGjRvlg5iYGNPpYWFh8sHBgweFEGvWrKmurq5ryWlpacbLR9atWycf9O7d26yZRqMZMGCAEGLz5s2lpaWOrrlWbm5uJ0+e9PHxsfwSjltfC4yrU7Pxk08+GRgYKIR45513DP/ZJAkhKioqPv300yeeeMLf3/+Oy3ecgICAXbt23bhx45lnnrGmvfFiLJsvgG2e3V5VVWV8rNU65DQxyx8QeSMPIUR5efkdF1VXx7Zo0UIIsWbNGtN7oGi12tOnT1tT4dChQ7t16yaE2LVrl/G8ShveD9aU0ZBNTePR2HPryJEjQojg4GDrz8GV94/RaDTBwcGm0318fOQHIzc3Vwghv87o9fpar1js3r17XFycfPzTTz/JlgEBATVbyi8+lZWVxvvWOK7mhnDc+lpgPH/SdBMpeXt7y0j45Zdftm7dapy+bt263Nxc5U63a9WqlXxQWFho2xKaZ7ebnk8kA7Uxq6tjJ0yYIITIycnp2bPn7373u++++66ysrJeS5bH7qqqqk6cOCGn2PB+sKYMJ2xqnKCx55Y8R8t4/NcaBQUFQgiDwSBPCjIlR1F+sZJLlt9QLMvLy7PQ0libbObQmi2LjIys9b5tMrEct742mzJlivyqazx4JYR4//33H3744XvuuaeBC3ey0NBQ+cB4Zlp9Nc9uNx54DAkJkVU5VGFh4cKFC4cNG9apU6fg4GAfHx+dTlfr+YH1MnLkyPT0dHkKz7Jlyx566KH27dvPnj375s2bVi6hffv28sGNGzfkAxveD9aUYZdNjcs19tySXWnDFZp6vb6uY6Pr1683NisoKJBHz2xWc3ZH19wQjlhfm7Vt21Z+Q/z222/leb0nT57cvXu3DZfvuNzAgQPlA+PPEvbVVLvd+LOK3OFwqMuXL8fFxc2cOdPb2/uTTz45cuRITk5OaWmpNV/m7mjGjBmnT5/++9//Lo/4ZWdnz5s3r2vXrvL74h0Z9+SsPFha1/vByjKcsKlxqMaeW/JQW3Z2tulxcMvkAfrS0lLLlzQFBQUJISorK+/4xmrdurX4vwc0TBUVFZk2c2jNlp06darWN+J9990nHLm+DTFr1iwhRHV19XvvvSeEeP/99++9994hQ4Y0fMlO1r9/f3nZVmZmZlZW1h3bX7lyxWxz2Ty73XhKgvHEFseZM2fOuXPnoqOj16xZEx8f3759e71eb8cf1Tp06PDCCy8cPXr0l19+GTdunBAiKytr5syZ1sxr3CUyHr6z+f1guQy7bGpcrrHnljxwUVZWdujQIStniYyMlA/kAeK6yIsKhRCmR/kttCwrK6v1EM3Zs2eFEDqdLjo62tE1N4Tj1rchunTp8vDDDwshPv7449zc3M8//3zatGkNX6zzeXp6/uUvfxFCVFdXv/vuu3dsn5mZaXbzlGbY7dnZ2Rs2bBBCdOjQYezYsY5+OXnGQWxsrKNvHt+5c+dVq1aNGTNGCGF2OX9d5Pi2aNHCuClo+Puh1jKcsKlxgsaeW/KqQCHEG2+8YeUsxvPl5s6da+HoSlJSknzw0ksvXb9+3cICExMT5YN9+/aZPWUwGHbt2iWEGDFihKenp6NrbgjHrW8Dye/+BQUFSUlJbm5ujz32mF0W20AFBQUDBgwIDAy0/n8bSUtLe/DBB4UQixcvlufmWFAzt5pht8+cOfPWrVtubm4ff/yxg04mNFVWViaEuHbtmt2XPHLkyJrny8g3g0ajuePsVVVVMlcSExONmWrD+8GaMpywqXGCxp5bycnJ8g6b33zzzfjx448dO1ZRUVFSUnL+/Pk9e/bUOsugQYN69OghhNi4cePQoUO3bdsmf9QpKirKzMxct26d3CXv1q2bvMdMVlZWz549ly5d+uuvv5aXl1dVVeXm5h46dMh4j+rY2Fj5nWXhwoVm5xGtWLEiKytLp9OZ3knFcTU3hOPWt4ESEhLksfjdu3c/9dRTdd1NwMk++eST3bt337hxY/r06Vb+VKnT6b788st+/frdvn176NChBw4cqKvlhQsXVq5cWXN/q/l0e2Vl5axZs2QnLFmyxPorRhri7rvvFkJs27Zt+fLlJSUlxcXFe/bsmTRpkjWXFlhmMBg+++wzs4nyZHR5wwvLli1bdvXqVb1e/9JLLxkn2vB+sKYMJ2xqnMHCtV320sD7Zezbt6/Wc/OM167XvPfE6dOnjZcjmAkNDc3Pz5fNioqKhg0bVlfPjB492rjA/Px8ed+mlJSU48ePl5eXX716NT09Xa/Xe3p6rlq1ymk113VFpDUct751VXX48GG58K+++spCYZ9//rkQQqvV3vHKUzNy4Y647ljeJFdYfb8Mo8rKyr///e+BgYFubm6pqalff/21vDNCUVHR2bNnV65cOX78eHniXIcOHczmbfzdXt/rjl977bW8vLzS0lJ5x4eCgoLMzMy33npLHq3q2bNcekihAAAgAElEQVTnvn37rHxpg63XHRt99dVXNd/2vXr1knk2bdq0Oy6kro4dPny4Tqd74YUXzpw5U15efvny5QULFri7uwcHB//73/82bSaEiIyMPH36dFFRUWVl5fXr1xcvXqzX6729vf/1r3+ZvVx93w9WlmHlpqa+uF+GuaysrLS0tE6dOnl5eXl7e3fv3n3u3LlFRUUyG7y9vfv27ZuTk2M6S2Fh4euvv96nT5/AwECtVhsUFJSQkPDuu+8WFxebNquurv7666+Tk5PDwsLkuyEsLCwpKWn16tVVVVWmLW/fvv3pp58OGTIkKChIq9X6+vrGxMRMnz793Llzzql5+/bt3bp1Mx5OCQoK6tu3744dO+rVk3Zf3yNHjsTGxhqrCgkJkTfok8fZjFew+fj4DBgwoKKiotaqKioq2rdvP27cOGtW4eTJkwkJCVFRUabfDPz9/bt06TJkyJBr167VOtfRo0eHDh0aFxfXoUMH4/866ObmFhQU1KVLl4EDBy5atMi0/Y0bN+p1vwwzxcXFn3322RNPPNGlSxfZgR4eHgEBAVFRUUOHDp05c+YXX3xx5cqVmjM2wm63vuvy8/NHjBgRHR3t5+cn73en1WpND5T5+PiEhob2799/9uzZu3fvtqYnrR9uaz4gS5cujYyM9PDwaN26dXx8/EcffVRZWSl/WtPpdN27d1+yZIkNHfvPf/5z4sSJXbt29ff3d3d3b9GiRZcuXWbNmnX16lXTV5e5FRwcLO97q9PpvL29o6Kipk+ffuHChVpXv16bHSvLMFi9eawXZ+aWxlDjykS7Gzt27Nq1a2fMmGG8VykANDcjRozYsGFDfHy8/HWqiTlw4IC8tYcTMqWx/74FAE1Jw39OA7kFAM6j7ll8jQe5BQBQCbkFAFAJuQUAUAm5BQCO9cUXX0RGRsr/++qHH34ICQnp37//0aNHXV2Xqhx+bxUAaOZSU1NTU1NdXUXTwf4WAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAlWqe90rVr1w4cOOC0lwMAOE1mZqbzXszgeMnJyc5bHwCA6zghUzhOCABQicZgMLi6BqCpWbhw4cyZM3v37r1//35X1wI0NexvAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFSidXUBgML27Nlz8uTJmtMzMjKEENevX//ggw9qPuvp6fnEE084vDigidIYDAZX1wCoatWqVampqfWda/To0V999ZUj6gGaA44TArb77W9/26JFi/rONX78eEcUAzQT5BZguxYtWowaNapes7Rs2XLkyJEOqgdoDsgtoEHqu/M0ZswYvV7voGKA5oDcAhpk6NChrVq1sr49BwmBBiK3gAbR6XRJSUlWNg4KCkpISHBoPUCTR24BDWX9LtTYsWO1Wi4+ARqE3AIaKj4+/q677rKmJQcJgYYjt4CGcnNzS0lJuWOz0NDQPn36OKEeoGkjtwA7sGZHKjU1VaPROKEYoGnjfhmAfdx777213vPJ6KeffuratavT6gGaKva3APsYO3ashWejoqIILcAuyC3APiZOnGjh2QkTJjitEqBpI7cA++jUqVOPHj3qetby3hgA65FbgN3UdXbGAw88EBER4eRigKaK3ALsZvz48e7u7rVOd34xQFNFbgF2ExIS0r9/f7OJbm5uycnJLqkHaJLILcCeau5aDRo0KCQkxCXFAE0SuQXYU3JysoeHh+kUDhIC9kVuAfYUEBAwZMgQ45+enp6JiYkurAdoesgtwM5Md7AeeeQRf39/FxYDND3kFmBno0eP9vHxkY85SAjYn6E2GRkZrq4LANCszZgxo9aEYn8LAKCSO/zXq+x4ATaorKwcMWLEK6+80qtXL1fXAqjnxRdf3LZtW13P3iG3+NQBtklLS0tLS9PpdK4uBFBPYGCghWc5Tgg4xJw5cwgtwBHILcAhzK4+BmAv5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAl5BYAQCXkFgBAJeQWAEAlzsutzz//PC4uztfXV6PRaDSaFStWOO2lbebamgcPHty2bVv50s8884wzX9qyV155pXPnzlqtVqPR+Pv7u7oc2BODa7MdO3b06tWrdevW8jM7ffp0V1fUZDkvtx5//PEff/zxxx9/dNorNpxra962bdu1a9datWrlkle34JVXXjl9+nSjilLYC4NrswcffPDAgQO5ubktWrRwdS1NHMcJAYfLy8sbOHDgb37zG7kfY0ar1fr5+UVERIwcOXLRokW5ubmurrceal21Tz75pGbLkydP9u/fPzQ0VLb09fXt2rXr66+/7vyaXeL06dNRUVEhISHfffdd8yzAjsgtwOFatWq1c+fOrKysF198UU55++23DQaDwWCorq4uKCjYu3fvU089lZGRMWPGjPDw8PXr17u2YOuZrZpGoxFCTJs27ezZs2Yto6Ki9uzZc+HChb/85S9CiHPnzh07duyvf/2r82t2idWrV586derq1atLlixpngXYEbkFuJJGo/Hx8YmJifnzn/+8Z88eT0/PoqKi1NTUCxcuuLo0W0ybNk0IUVxcPGHChMrKSleX04iMHTs2IiIiODj46aefbp4F2BG5pQb5NdbunnvuuUcffdQRS4YNoqKiRo0aJYQoLS1ds2ZNQxblqpH985//3K9fPyHEoUOH/va3vzm/gEYrMjLy9OnT169fHzJkiKNfq9bRd2YBjubs3HJzUy8pG0PN7u7udl/m9evX33333QYuxEGB2myFh4fLB7/++qvNC7HLyAqbBtfNzW3ZsmU+Pj5CiDfeeGPv3r0NLwP1Yq/Rb8ycvUXW6XROfsWGaww1a7Vauy9z3rx5t27dauBCHBGozVleXp580JBz0O0yssLWwb377rvT09OFEFVVVY899lhRUVHDK4H17DX6jZkDc6uqqmrZsmXDhg1r27ath4eHr69veHj4448/Xlf74uLiefPm3X///f7+/h4eHqGhoRMnTjx69Gitjbdv3/7YY4+Fh4d7e3t7eXm1bds2Ojo6MTFxx44dpgWsWLFixIgRISEhnp6efn5+MTEx06ZNO3funNNq3rFjR7du3XQ6nUaj6dmzpxCiurr6yy+/fOSRRwICAoqLiy33oZGXl9f+/fsTExODgoI8PT3bt2+fmJi4devWmi1Pnjz56KOPduvWrV27dnq93sfHJyIi4qmnnrpy5YqxzeLFi6OjoxcvXiyEWLdunTwHzGxDaU0PCyE8PT2FEHv37n322WfvuecevV7fsWPHp59+urCw0LTZ0aNHY2NjZT/ExsYKITIyMiZPnty5c2e9Xh8eHj5lypSSkpJa193KcbTc1Zs3b46NjfXw8NBoNJ07dzYYDNu2bZs0aVJERISXl1f79u2feeYZeSLfyZMnX3311fj4+FatWvn6+vbu3XvDhg01qyosLBw4cGBgYOCHH35oYeDq6+TJk0IIjUYjDxjWfNby4NpxZIXVg1vTH//4x+HDhwshsrKynn322Xr1QOMcbms+VhYkJSWFhITI4fjTn/5knD5p0qSa55cade3atV41WBj9ugqob7c35INsT4baZGRkWHjWGnl5eX379hVCDB48+ODBg6WlpaWlpdnZ2d98841c8vLly03bnz9/vnPnzkKIqVOnXrp0qaSkZOXKlXq9XqfTrV+/3rRlRUVFamqqEKJXr17bt28vKSnJzc3dv39/WlqaEGL+/PmyWX5+fv/+/YUQ48ePP378eFlZ2ZUrV9LT0729vT09PVeuXOnMmmVt4eHhW7du7dKli7Hzb968eceelNdvRURE1HrQZtasWWbtN27c6ObmtmDBghMnTpSUlJw9ezYhIUEIERYWVlhYaNpSVpKUlGS2BCt7WP4ZGBjYo0ePmoUNHTq05rrIWfz8/Pr06VNzlpqVGOo/jpa7etasWUIILy+v6OjomgV0795dvpYZrVZ7+PBhsxeSuxRyaeXl5XcYxf+YM2eOnMt4PqGpbdu2yVGeMmVKrbNbObgNHFmDTYMrV+3q1avyz6tXr7Zu3Vq2/+KLL0xbvvDCC0KInJycmgtptMNt/cfKYDDI67emTZtmNl1+ltPS0oxT/vCHP/Tp02f//v2mC0lKSpKVbNy40YYa6hr9WgtoSLfX64NcX8nJyUKIGTNm1Pqso3JLftuKi4sz+0ifOXNGLtksA+RvuaNGjTKdKH/XDQgIMB2VyZMnCyGioqJu3bpl9qIRERHGz97o0aOFEAkJCdXV1aZtVq1aJYTQ6XQ135qOq1kOs06nkxfobNmyRX4nsj63+vTps3PnzpycnMrKymvXrn3wwQd+fn61bhQ2btz44IMPmk755ZdfZMsPP/zQdHpd728re1iulI+Pz/Hjx7Ozs2XM79mzx/gLzalTp8xmN86SmZmZk5NTVlZWUlKyZcuWkJAQOcvFixfNZqnvOFruarkh02q1a9euPXHixM2bN0tKSnbs2NGxY0dZwAsvvLBly5asrKzy8vKrV6/OnDlTTv/jH/9oVtiCBQvkU56eng3MrZKSkp9//vmVV17R6/UBAQFvvvmm2coaWTm4DRxZg02Da5ZbBoPhyy+/lI39/PyysrKM0y3kVqMdbus/VoZ65tb7779v2uZf//qXXGxKSoptNdiQW7Z1e70+yPXlgtwy/hi7du1as6dqzYDdu3fLiZs2bTJtfPz4cTl92bJlcsqJEyfkWRLGKaYmTZokP3vff/+9nHHDhg1mbaqrq+Vnb/jw4c6p2WDy9cQ4paqqKjIy0vrcevrpp82m79q1S75Q165dLS+hurpa/kT33HPPmU6v9f1tZQ/XulLSunXrZGHr1q0ze6quWZYuXSpn2bx5s+l0G8bRclfLDZmnp6fZ0oxb2B07dphOr6qqCg0NFUL079/fbJb8/Pz4+Hh/f/8lS5YYrGbMrVq1bdt2wYIF33//vfULrHVwGziyBpsGt2ZuGQyGiRMnyvb9+/evqqqSE+vKrcY83DXV9bEy1DO3Dh06ZPyzpKQkLCxMrpFZT1pfQ31zyy7dLtX1QbaB5dxyyO9ba9eulQ8eeugha9pv3LhRPoiJiTGdLodQCHHw4EH5YM2aNdXV1XUtOS0tbdCgQUII46erd+/eZm00Gs2AAQOEEJs3by4tLXVCzbVyc3M7efKkPO3KNvHx8fKo5rFjx7Kzsy201Gg0Hh4eQojy8vI7LtbKHrbAeLjGtHstM/ah2Sw2jGNN1nS1sebLly+bzSsPBdd8iYCAgF27dt24ccO2WyKZ7m9VVFTk5OTs27fvmWeemTdv3n333de1a9cDBw5YsxzrB7fhIyvqP7jvvPNOhw4dhBB79uyZN2+e5caNebhrqtfHqi4fffSR/HFOmjNnzvnz54UQ8+bNa9u2rXNqsEu3S3V9kO3OIbl15MgRIURwcLDxWJZlmZmZQgiNRhMcHGw63cfHR55HZ7zzjfxqoNfr27VrV3M53bt3j4uLE0L89NNPsllAQEDNZvIbRGVlpXxdR9fsOMaj86dPn7bXMq3sYQuMJ6EZDAYrX9R4xqbZLDaMo22Mp2vWrFmujvXrYgOdTte6des+ffrMmTMnMzPznnvu+fnnnx988MF9+/bZ8VUaPrKi/oPr5+f36aefyh/tXnnllcOHD1to3EyGuy7Hjh1btGiREKJ3797OvDTYjt1e1wfZ7hySWzk5OUIIX19fK9sXFBQIIQwGgzxNxZS85N74bUIu+Y63rZQnE9fVzFiY8Zxjh9ZsWWRkZK2nEhl33i1o06aNfGB6fldhYeHChQuHDRvWqVOn4OBgHx8fnU5n/Rk+Vvawc9gwjqoLDg6eP3++EKKsrEzee8JUQwbXVSM7ePDgKVOmCCFu3749YcIEC9U25uFu4Mfqjqqrq5966qnKykqtVrt06dJaT8JyUA2NudvrYv+rgoQQcsNdUVFRr7n0er2Vlx0UFBRUV1fbfDmwPFpixtE1O4Lxy69er5cPLl++HB8ff+7cucTExE8++SQ8PDwwMFCn0/n7+9fr/d3AHnaOWsexCTAer/vhhx9yc3ONZ+XZZXBdMrLz5s3bsmXLqVOnzpw5M336dOP3rXpx4XDb62Nlwfvvvy9/WZg1a5bZTw9Oq6FWjfNT5pDckofasrOzq6qqrLl0UV5hUFpaWlxcbPnAdFBQkBCisrIyJyfHwrtfftTrujrKeCGkcYvg0JotO3XqlM3zGnezjEd+5syZc+7cuejo6DVr1th20aiVPewcNoxjE9CiRQt3d/eqqiohRElJiXHtGji4LhxZvV6/fPny3r17V1ZWfvTRR927d6+1WaMd7oZ/rCy7du2avL9wx44d6zpzx3E1NNput8AhX7vuueceIURZWdmhQ4esaR8ZGSkfyCOtFsgr3YQQtV5ya9asrKys1n1beadqnU5nelWH42p2HHmiY8uWLWXx4j8ng8TGxtr8zrayh53DhnFsAuSXJyGEp6en6Y/zDRxc147sfffdZ7wXfl03E2i0w93wj5Vl06ZNk99B33vvPeOxEyGEwWAwnivhuBoabbdb4JDcklfDCSHeeOMNa9obT3CaO3eu5d1S4xV5L7300vXr1+tqlpiYKB/U/GXb8J8zyEeMGCFvB+Domh2kurp6+/btQoixY8caD/uUlZUJIa5du2bNEuQxdLPvWVb2sHPYMI5OU1BQMGDAgMDAQOO5v/ayZcsW+WDo0KGmq2b94DbOkX3hhRdMz52rqdEOd70+VvW1adMmeQ/llJSUhx9+2PSp48ePG3/oauDoW9Bou90Ch+RWcnLyXXfdJYT45ptvxo8ff+zYsYqKipKSkvPnz+/Zs6dm+0GDBsnL8jdu3Dh06NBt27bJo/BFRUWZmZnr1q27efOmbNmtW7eUlBQhRFZWVs+ePZcuXfrrr7+Wl5dXVVXl5uYeOnToH//4hxAiNjZ2zJgxQoiFCxeandmyYsWKrKwsnU738ssvO6dmB/n0008vXbrk6+truiJ33323EGLbtm3Lly8vKSkpLi7es2fPpEmTaj0tVR4s2rVr1+7duysqKq5evXr9+nUre9g5bBhHp/nkk092795948aN6dOn1/dnUQuys7NfeuklIYSPj4/Z/6lo/eA2zpHVarXLly833Z8w02iHu14fq3opLS2V98Hy9/eXJxOa2r17t/G/O2/g6FuoodF2uyW1XtXV8Ptl7Nu3r9Zz84wXVJvde+L06dPGK5/MhIaG5ufnG1sWFRUNGzasrtUZPXq0bJafny8vb0pJSTl+/Li8Hj49PV2v13t6eq5atcqZNdd1mZ415M8SSUlJFy9eLC0traqqunz58vz58728vPz9/Xfu3Gna+KuvvqpZTK9eveSb3uxCSHnemin5xrWyh+taqbpuL2JhFuPp0V999ZXZU/UdR8tdXdeFqBZqHjp0qBCiR48eZtPffPNNOYtt98t44403CgsLjTPeunXr1KlTixYtkm+28PDwffv2mc1r/eA2cGQtdKOFjqr1uuOa3nrrLbmEuu7z1AiHu14fK0N9rjt+/vnn5dJMr16vrKy8cOHC8uXLIyIinnjiifrWUNfo11qAZK9ut/BBri/X3OfJYDBkZWWlpaV16tTJy8vL29u7e/fuc+fOLSoqktng7e3dt29f0/duYWHh66+/3qdPn8DAQK1WGxQUlJCQ8O677xYXF5stubq6+uuvv05OTg4LC5M9GxYWlpSUtHr1auM1+QaD4fbt259++umQIUOCgoK0Wq2vr29MTMz06dPPnTvntJq3b9/erVs34/UiQUFBffv2NbtK37KrV6++9957qampXbt29ff3l/+ne2xs7F//+tfs7Oya7ZcuXRoZGenh4dG6dev4+PiPPvqosrJy7NixQgidTte9e/eMjAzZsqysbMqUKW3atPHy8oqKinr22WeNN+Ox3MNmKxUSEvL73//eYDBUVFQMGjTIeBWIt7d3v379CgoKDAbDkSNHYmNja84ij7YZr5nz8fEZMGBARUWF6RpZOY6Wu3rv3r3yfqDy2TZt2gwaNKigoCAzM7Nv377GO8/q9frY2NjVq1cbDIbJkyfffffd8qiLRqMJDw83fv4NBsONGzesv19GXl7e4MGDO3bsWNd9/d3d3f39/aOiolJTU1evXl1WVlbrcqwcXJtHtmY33nFwzVatZcuWMTExf/rTn+rqiurqanlMvtbcarTDbU3P79ixo0+fPsbzF9zd3cPDw4cNG7Zv376+ffvKL6By+t133/3www9fvHjxjv/XxMyZMxs++l988UWtBdiwtWzIB7leLOeWxlDbBWIHDhyQvwfW+iwAAI4zduzYtWvXzpgxw3gDa1ON+gIdAADMkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlZBbAACVkFsAAJWQWwAAlWgtP63RaJxTBwAA1mB/CwCgEo3BYKg5tbi4+Pjx486vBmgaVq9evXjx4ujo6A8//NDVtQCqateuXVhYWM3ptR8n9PHx6dWrl4NLApqsjIwMIUTLli35HAF2x3FCAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKtK4uAFBYVlZWXl5ezemXLl0SQpSUlPzwww81n9VoNHFxcQ4vDmiiNAaDwdU1AKp6++23p06dWt+5+vfvv3v3bkfUAzQHHCcEbJeSkqLV1vugxfjx4x1RDNBMkFuA7YKDgwcNGlSvWbRabVJSkoPqAZoDcgtokPruPA0ZMiQ4ONhBxQDNAbkFNEhiYqJer7e+PQcJgQYit4AG8fX1HTZsmJWNvb29R48e7dB6gCaP3AIayvpdqJEjR/r4+Di0GKDJI7eAhhoxYoSfn581LTlICDQcuQU0lJeXlzVH//z9/R9++GEn1AM0beQWYAfW7Eg9+uijnp6eTigGaNrILcAOEhIS2rRpY7kNBwkBuyC3ADtwd3dPTk620KBdu3YDBgxwWj1AE0ZuAfZheXcqJSXF3d3dacUATRj31QXsw2AwhIeH//rrr7U+e/Dgwfvvv9/JJQFNEvtbgH1oNJpx48bV+lR4eHjPnj2dXA/QVJFbgN3UdagwNTVVo9E4uRigqSK3ALuJiYmJjo6uOT0lJcX5xQBNFbkF2FPNXa7u3bt36dLFJcUATRK5BdjThAkTzA4JctkWYF/kFmBPYWFhvXr1Mv5p4WQNALYhtwA7Mw2qfv36hYaGurAYoOkhtwA7GzdunFarlY85SAjYHdcdN2u9e/d2dQlN08mTJwsLCzUaTVxcnE6nc3U5TdDcuXMHDRrk6irgGlpXFwBXOnDggKtLaMoMBsMPP/zg6iqapry8PFeXAJchtyCmTZsWFhbm6iqalLKysjlz5owdO7ZHjx6urqWpefXVVwsKClxdBVyJ3IIYN26c6SlwsIvjx4+/9957vr6+ri6kqUlPTye3mjnOywAcYtGiRYQW4AjkFuAQQUFBri4BaJrILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgtAIBKyC0AgErILQCASsgt2NPnn38eFxfn6+ur0Wg0Gs2KFStcXdGdqVgz0JyRW7Cnxx9//Mcff/zxxx9dXUg9qFgz0JxpXV0AAHN5eXlJSUlZWVmXLl2qqqoye1an0/n6+gYGBnbp0iUuLi4xMbFLly4uqRNwCfa3gEanVatWO3fuzMrKevHFF+WUt99+22AwGAyG8vLyCxcurF+//sknnzx16tTLL78cHR3du3fvH374wbU1A05DbgEq8fDwaNu2bb9+/f7yl7+cOHFi5cqVvr6+Bw4ceOCBB5YuXerq6gBnILcAVWk0mtTU1L1797Zs2bKqqurZZ5/9xz/+4eqiAIcjt2B/bm7qva9UrFmKiYn54IMPhBDV1dVTp069deuWqysCHEvVzyoaM51O5+oS6k3Fmo1SUsETjA0AAApFSURBVFJiYmKEEJcvX/7ss89cXQ7gWOQWGqSqqmrZsmXDhg1r27ath4eHr69veHj4448/bmGW4uLiefPm3X///f7+/h4eHqGhoRMnTjx69GjNltu3b3/sscfCw8O9vb29vLzatm0bHR2dmJi4Y8cO0wJWrFgxYsSIkJAQT09PPz+/mJiYadOmnTt3zmk179ixo1u3bjqdTqPR9OzZUwhRXV395ZdfPvLIIwEBAcXFxUKIwsLCgQMHBgYGfvjhhxZ71BYajSYlJUU+/vbbb+tV/NGjR2NjY2XxsbGxQoiMjIzJkyd37txZr9eHh4dPmTKlpKSk5ouWlJTMnz+/R48egYGBLVq06NixY0JCwssvv1yvVwdsYUAzJt8DGRkZts2el5fXt29fIcTgwYMPHjxYWlpaWlqanZ39zTffyCUvX77cbJbz58937txZCDF16tRLly6VlJSsXLlSr9frdLr169cbm1VUVKSmpgohevXqtX379pKSktzc3P3796elpQkh5s+fL5vl5+f3799fCDF+/Pjjx4+XlZVduXIlPT3d29vb09Nz5cqVzqxZ1hYeHr5161bTE9Nv3rxpMBjS09Pln15eXuXl5Vb28Jw5c+RcxvMJ67J3717Z0tfX1+bi/fz8+vTpU3MrkZSUZPZyhYWFUVFRGo1m8eLFOTk5ubm5mzdvjo2NbdWqlQ2vXi933XWXEGLNmjW2zY4mgNxq1hqYW8OHDxdCxMXFmW2Iz5w5U1cG9OvXTwgxatQo04l/+9vfhBABAQGFhYVyyuTJk4UQUVFRt27dMltCRESEMbdGjx4thEhISKiurjZts2rVKiGETqc7fPiw02qWm36dTjdy5MhFixZt2bJl8eLFxtxasGCBXL6np6cjcuvChQvGmCktLbWteB8fn8zMzJycnLKyspKSki1btoSEhMhlXrx40XQJr776qhBizJgxphN37txplltWvnq9kFsgt5q1huSW8Qv+2rVrzZ6qKwN2794tp2/atMl0+vHjx+X0ZcuWGQyGEydOyLMk5J9mJk2aJHPr+++/l3Nt2LDBrE11dXV4eLgQYvjw4c6p2WCyy2JsU1VVFRkZKXMrPz8/Pj7e399/yZIlNVeqLtbnlumhPGPGNKR4yXhu/ebNm02njxgxQgiRlpZmOvH27dsRERHGP61/9Xoht8DvW7DR2rVr5YOHHnrIylk2btwoH8iTCIzCwsLkg4MHDwoh1qxZU11dXdeS09LSBg0aJIRYt26dnNK7d2+zNhqNZsCAAUKIzZs3l5aWOqHmWrm5uZ08edLHx0cIERAQsGvXrhs3bjzzzDNWvnS9mN5WQ6v93/vgNKR4sxlNu1EI0aJFCyHEmjVrfv75Z9PXPX36tPHPhr86UCtyCzY6cuSIECI4ONjPz8/KWTIzM4UQGo0mODjYdLqPj4/c1Obm5goh5I6UXq9v165dzYV07949Li5OCPHTTz/JZgEBATWbyf2tyspK+aKOrtnl5NkfUmBgoHzQ8OKNp1ka/rN3Lk2YMEEIkZOT07Nnz9/97nffffddZWWl2byqdB2UQ27BRjk5OUIIX19f62cpKCgQQhgMBnn2mim51SsvLzcuWX6jtyAvL89CM2Nhspmja3a5ixcvygchISEeHh7yseOKHzlyZHp6ujzHZNmyZQ899FD79u1nz5598+ZNYxtVug7K4b66sJHc9FRUVNR3Rr1eb821sQUFBdXV1TZfDiyPNJpxdM0uZPzRSJ4KYcpBxc+YMePRRx/9/PPP165d+9NPP2VnZ8+bN2/16tWHDh0KCgpy9KujOWN/CzaSh9qys7Nr3rC8Lv7+/kKI0tJS04NaNcmtXmVlpdw9qkvr1q3F/z0+ZqqoqMi0maNrdrlNmzbJB6NGjTJOdHTxHTp0eOGFF44ePfrLL7+MGzdOCJGVlTVz5kznvDqaLXILNrrnnnuEEGVlZYcOHbJylsjISPlA/jRVF3n1qxBi69atd2xWVlZmeiTQ6OzZs0IInU4XHR3thJpdKzs7e8OGDUKIDh06jB071jjdacV37tx51apVY8aMEUIYrwpXouugInILNkpISJAP3njjDStnMZ7FN3fu3FqP40lJSUnywUsvvXT9+vW6miUmJsoH+/btM3vKYDDs2rVLCDFixAhPT08n1HxHBQUFAwYMCAwMdMRd22fOnHnr1i03N7ePP/7YeDKhsF/xNY0cOdL0hBfpwQcfFEJoNBpHvzqaOXILNkpOTpZX0nzzzTfjx48/duxYRUVFSUnJ+fPn9+zZU+ssgwYN6tGjhxBi48aNQ4cO3bZtm/wRq6ioKDMzc926dfJX/W7dusm7FmVlZfXs2XPp0qW//vpreXl5VVVVbm7uoUOH5F3PY2Nj5Rf8hQsXmp3ttmLFiqysLJ1OZ3bbIcfVfEeffPLJ7t27b9y4MX36dBt+YKtLZWXlrFmzVq5cqdPplixZYnZ+v72Kr8lgMNS8EaI8J17ewcShr47mznWXjsH15HvA5vtl7Nu3r9Zz84w3Wah574nTp08bL98xExoamp+fL5sVFRUNGzasrjft6NGjZbP8/Hx506aUlJTjx4+Xl5dfvXo1PT1dr9d7enquWrXKmTXXdemu9Oabb8pZbLtfxmuvvZaXl1daWirvDFJQUJCZmfnWW2/JY3E9e/bct29frUtoYPGHDx+Wjb/66ivT6cOHD9fpdC+88MKZM2fKy8svX768YMECd3f34ODgf//73/V99XrhumOQW82a3ILYnFsGgyErKystLa1Tp05eXl7e3t7du3efO3duUVGRzAZvb+++ffvm5OSYzlJYWPj666/36dMnMDBQq9UGBQUlJCS8++67xcXFps2qq6u//vrr5OTksLAwmUNhYWFJSUmrV6+uqqoyNrt9+/ann346ZMiQoKAgrVbr6+sbExMzffr0c+fOOa3m7du3d+vWzXiALigoqG/fvjt27DBdwo0bN6y/X0Z+fv6IESOio6P9/Pzk3fy0Wq3x+JsQwsfHJzQ0tH///rNnz969e7flpVku/siRI7GxscbiQ0JCfv/73xsMBnlg03ihm4+Pz4ABAyoqKuQy//nPf06cOLFr167+/v7u7u4tWrTo0qXLrFmzrl69Wq9XtwG5BY3h/x5gQbMiN4UZGRm9evVydS2AVTp06HDp0qU1a9YkJye7uha4Br9vAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFSidXUBcL0//vGPPj4+rq4CsEp2drarS4CLkVsQx48fd3UJAGAtcqtZS09Pd3UJgC26d+/u6hLgMhqDweDqGgAAsBbnZQAAVEJuAQBUQm4BAFRCbgEAVEJuAQBUQm4BAFRCbgEAVPL/AI8hBii+54VLAAAAAElFTkSuQmCC\n","text/plain":["<IPython.core.display.Image object>"]},"metadata":{"tags":[]},"execution_count":16}]},{"cell_type":"code","metadata":{"id":"RWXO02XtgVHm","colab":{"base_uri":"https://localhost:8080/","height":1000},"executionInfo":{"status":"ok","timestamp":1603800944465,"user_tz":240,"elapsed":375162,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"8227eff9-3440-420d-812f-f0131741e52e"},"source":["autoencoder.compile(loss='mean_squared_error', \n","                    optimizer = 'adam')\n","\n","history = autoencoder.fit(x=X_train_y0_scaled, \n","                          y=X_train_y0_scaled, \n","                          epochs=100, \n","                          batch_size=128, \n","                          validation_data=(X_valid_y0_scaled, \n","                                           X_valid_y0_scaled),\n","                          verbose=1).history"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/100\n","101/101 [==============================] - 4s 38ms/step - loss: 3.1838 - val_loss: 1.6138\n","Epoch 2/100\n","101/101 [==============================] - 4s 35ms/step - loss: 1.0715 - val_loss: 1.2308\n","Epoch 3/100\n","101/101 [==============================] - 4s 35ms/step - loss: 0.8191 - val_loss: 1.0489\n","Epoch 4/100\n","101/101 [==============================] - 4s 35ms/step - loss: 0.6967 - val_loss: 0.8912\n","Epoch 5/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.6506 - val_loss: 0.8225\n","Epoch 6/100\n","101/101 [==============================] - 4s 35ms/step - loss: 0.6025 - val_loss: 0.6632\n","Epoch 7/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.5654 - val_loss: 0.5733\n","Epoch 8/100\n","101/101 [==============================] - 5s 50ms/step - loss: 0.5329 - val_loss: 0.5548\n","Epoch 9/100\n","101/101 [==============================] - 4s 41ms/step - loss: 0.4983 - val_loss: 0.5214\n","Epoch 10/100\n","101/101 [==============================] - 4s 39ms/step - loss: 0.4902 - val_loss: 0.5255\n","Epoch 11/100\n","101/101 [==============================] - 4s 38ms/step - loss: 0.4712 - val_loss: 0.4856\n","Epoch 12/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.4453 - val_loss: 0.4486\n","Epoch 13/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.4449 - val_loss: 0.4531\n","Epoch 14/100\n","101/101 [==============================] - 4s 35ms/step - loss: 0.4065 - val_loss: 0.4286\n","Epoch 15/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.4078 - val_loss: 0.4632\n","Epoch 16/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.3917 - val_loss: 0.3834\n","Epoch 17/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.3636 - val_loss: 0.4001\n","Epoch 18/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.3514 - val_loss: 0.3726\n","Epoch 19/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.3421 - val_loss: 0.3574\n","Epoch 20/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.3360 - val_loss: 0.3485\n","Epoch 21/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.3509 - val_loss: 0.3316\n","Epoch 22/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.3137 - val_loss: 0.3413\n","Epoch 23/100\n","101/101 [==============================] - 4s 35ms/step - loss: 0.3258 - val_loss: 0.3372\n","Epoch 24/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.3133 - val_loss: 0.3309\n","Epoch 25/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2901 - val_loss: 0.3042\n","Epoch 26/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.2838 - val_loss: 0.3130\n","Epoch 27/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2786 - val_loss: 0.2976\n","Epoch 28/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2764 - val_loss: 0.3240\n","Epoch 29/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2713 - val_loss: 0.2898\n","Epoch 30/100\n","101/101 [==============================] - 4s 38ms/step - loss: 0.2624 - val_loss: 0.3041\n","Epoch 31/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.2897 - val_loss: 0.3052\n","Epoch 32/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2594 - val_loss: 0.2949\n","Epoch 33/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2878 - val_loss: 0.3111\n","Epoch 34/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.2591 - val_loss: 0.2814\n","Epoch 35/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.2436 - val_loss: 0.2731\n","Epoch 36/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.2380 - val_loss: 0.2498\n","Epoch 37/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2266 - val_loss: 0.2711\n","Epoch 38/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2262 - val_loss: 0.2577\n","Epoch 39/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2176 - val_loss: 0.2632\n","Epoch 40/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.2146 - val_loss: 0.2517\n","Epoch 41/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2140 - val_loss: 0.2485\n","Epoch 42/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.2087 - val_loss: 0.2649\n","Epoch 43/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.2058 - val_loss: 0.2409\n","Epoch 44/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1975 - val_loss: 0.2435\n","Epoch 45/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1961 - val_loss: 0.2507\n","Epoch 46/100\n","101/101 [==============================] - 4s 38ms/step - loss: 0.1842 - val_loss: 0.2186\n","Epoch 47/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1867 - val_loss: 0.2300\n","Epoch 48/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1841 - val_loss: 0.2266\n","Epoch 49/100\n","101/101 [==============================] - 4s 35ms/step - loss: 0.1777 - val_loss: 0.2466\n","Epoch 50/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1750 - val_loss: 0.2639\n","Epoch 51/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1750 - val_loss: 0.2295\n","Epoch 52/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1785 - val_loss: 0.2400\n","Epoch 53/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1769 - val_loss: 0.2304\n","Epoch 54/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1671 - val_loss: 0.2004\n","Epoch 55/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1619 - val_loss: 0.2245\n","Epoch 56/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1591 - val_loss: 0.2910\n","Epoch 57/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1779 - val_loss: 0.2312\n","Epoch 58/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1562 - val_loss: 0.2020\n","Epoch 59/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1523 - val_loss: 0.2099\n","Epoch 60/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1505 - val_loss: 0.2349\n","Epoch 61/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1575 - val_loss: 0.2706\n","Epoch 62/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1496 - val_loss: 0.2288\n","Epoch 63/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1451 - val_loss: 0.2045\n","Epoch 64/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1349 - val_loss: 0.2016\n","Epoch 65/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1336 - val_loss: 0.2315\n","Epoch 66/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1386 - val_loss: 0.2100\n","Epoch 67/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1339 - val_loss: 0.2162\n","Epoch 68/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1240 - val_loss: 0.2542\n","Epoch 69/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1343 - val_loss: 0.3203\n","Epoch 70/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1357 - val_loss: 0.2205\n","Epoch 71/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1323 - val_loss: 0.2450\n","Epoch 72/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1251 - val_loss: 0.2396\n","Epoch 73/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1335 - val_loss: 0.2014\n","Epoch 74/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1183 - val_loss: 0.2526\n","Epoch 75/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1186 - val_loss: 0.2231\n","Epoch 76/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1130 - val_loss: 0.2372\n","Epoch 77/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1121 - val_loss: 0.2117\n","Epoch 78/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1185 - val_loss: 0.2278\n","Epoch 79/100\n","101/101 [==============================] - 4s 35ms/step - loss: 0.1226 - val_loss: 0.3078\n","Epoch 80/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1192 - val_loss: 0.2393\n","Epoch 81/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1107 - val_loss: 0.2899\n","Epoch 82/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1127 - val_loss: 0.2451\n","Epoch 83/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1105 - val_loss: 0.2823\n","Epoch 84/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1110 - val_loss: 0.2150\n","Epoch 85/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1102 - val_loss: 0.2354\n","Epoch 86/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1142 - val_loss: 0.2394\n","Epoch 87/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1123 - val_loss: 0.2273\n","Epoch 88/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1038 - val_loss: 0.2158\n","Epoch 89/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1015 - val_loss: 0.2556\n","Epoch 90/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1014 - val_loss: 0.2994\n","Epoch 91/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.0997 - val_loss: 0.2468\n","Epoch 92/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.0935 - val_loss: 0.2156\n","Epoch 93/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.0964 - val_loss: 0.2390\n","Epoch 94/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.0986 - val_loss: 0.2860\n","Epoch 95/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1012 - val_loss: 0.3316\n","Epoch 96/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1099 - val_loss: 0.3617\n","Epoch 97/100\n","101/101 [==============================] - 4s 36ms/step - loss: 0.1066 - val_loss: 0.2745\n","Epoch 98/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.0930 - val_loss: 0.2459\n","Epoch 99/100\n","101/101 [==============================] - 4s 37ms/step - loss: 0.1045 - val_loss: 0.2456\n","Epoch 100/100\n","101/101 [==============================] - 4s 38ms/step - loss: 0.0998 - val_loss: 0.2621\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"9xT4bX6pS3d8"},"source":["autoencoder.save_weights('cnn-sparse-autoencoder.h5')"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"0K-TgytePiVb","colab":{"base_uri":"https://localhost:8080/","height":631},"executionInfo":{"status":"ok","timestamp":1603789306662,"user_tz":240,"elapsed":378953,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"1413a828-b4e1-4bf0-c825-9503eabdce34"},"source":["error_vs_class_valid = rp.reconstructionerror_vs_class(model=autoencoder, \n","                                                       sample=X_valid_scaled, \n","                                                       y=y_valid)\n","\n","plt, fig = rp.error_boxplot(error_vs_class=error_vs_class_valid)\n","fig.savefig('cnn-sparse-autoencoder-master-boxplot.pdf', bbox_inches='tight')\n","\n","plt, fig = rp.error_boxplot(error_vs_class=error_vs_class_valid, ylim=[0.,1])\n","fig.savefig('cnn-sparse-autoencoder-zoomed-boxplot.pdf', bbox_inches='tight')"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"id":"obYIWCk2Pi_-","colab":{"base_uri":"https://localhost:8080/","height":34},"executionInfo":{"status":"ok","timestamp":1603789306665,"user_tz":240,"elapsed":378945,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"593934b8-51e8-4603-b219-989517c69c11"},"source":["errors_valid0 = rp.class_errors(error_vs_class=error_vs_class_valid, \n","                                label=0)\n","errors_valid1 = rp.class_errors(error_vs_class=error_vs_class_valid, \n","                                label=1)\n","rp.robust_stat_difference(z1=errors_valid0, z2=errors_valid1)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["2.964472357872913"]},"metadata":{"tags":[]},"execution_count":18}]},{"cell_type":"code","metadata":{"id":"2koGfbivQITq","colab":{"base_uri":"https://localhost:8080/","height":271},"executionInfo":{"status":"ok","timestamp":1603789308395,"user_tz":240,"elapsed":380663,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"d7112999-c64f-429e-e0c4-547619d72c55"},"source":["error_vs_class_test = rp.reconstructionerror_vs_class(model=autoencoder, \n","                                                      sample=X_test_scaled, \n","                                                      y=y_test)\n","conf_matrix, fig = rp.model_confusion_matrix(\n","    error_vs_class=error_vs_class_test, \n","    threshold=errors_valid1.quantile(0.50))\n","fig.savefig('cnn-sparse-autoencoder-baseline-conf-matrix.pdf', bbox_inches='tight')"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 288x288 with 2 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"eu-lSiUMluM0"},"source":["# Convolutional Encoded Classifier"]},{"cell_type":"markdown","metadata":{"id":"_ytb9sekfDHH"},"source":["## Encoding Model"]},{"cell_type":"code","metadata":{"id":"X8e-2tRybxVK","colab":{"base_uri":"https://localhost:8080/","height":609},"executionInfo":{"status":"ok","timestamp":1603800953964,"user_tz":240,"elapsed":1908,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"722bbf46-a1ea-491d-ed4d-ae3d8ae586db"},"source":["encoder_model = Model(inputs=autoencoder.inputs, \n","                      outputs=autoencoder.get_layer('encoder').output)\n","X_train_y0_encoded_features = pd.DataFrame(dp.flatten(encoder_model.predict(X_train_y0_scaled)))\n","X_train_y0_encoded_features"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>0</th>\n","      <th>1</th>\n","      <th>2</th>\n","      <th>3</th>\n","      <th>4</th>\n","      <th>5</th>\n","      <th>6</th>\n","      <th>7</th>\n","      <th>8</th>\n","      <th>9</th>\n","      <th>10</th>\n","      <th>11</th>\n","      <th>12</th>\n","      <th>13</th>\n","      <th>14</th>\n","      <th>15</th>\n","      <th>16</th>\n","      <th>17</th>\n","      <th>18</th>\n","      <th>19</th>\n","      <th>20</th>\n","      <th>21</th>\n","      <th>22</th>\n","      <th>23</th>\n","      <th>24</th>\n","      <th>25</th>\n","      <th>26</th>\n","      <th>27</th>\n","      <th>28</th>\n","      <th>29</th>\n","      <th>30</th>\n","      <th>31</th>\n","      <th>32</th>\n","      <th>33</th>\n","      <th>34</th>\n","      <th>35</th>\n","      <th>36</th>\n","      <th>37</th>\n","      <th>38</th>\n","      <th>39</th>\n","      <th>40</th>\n","      <th>41</th>\n","      <th>42</th>\n","      <th>43</th>\n","      <th>44</th>\n","      <th>45</th>\n","      <th>46</th>\n","      <th>47</th>\n","      <th>48</th>\n","      <th>49</th>\n","      <th>50</th>\n","      <th>51</th>\n","      <th>52</th>\n","      <th>53</th>\n","      <th>54</th>\n","      <th>55</th>\n","      <th>56</th>\n","      <th>57</th>\n","      <th>58</th>\n","      <th>59</th>\n","      <th>60</th>\n","      <th>61</th>\n","      <th>62</th>\n","      <th>63</th>\n","      <th>64</th>\n","      <th>65</th>\n","      <th>66</th>\n","      <th>67</th>\n","      <th>68</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>-0.001331</td>\n","      <td>0.001008</td>\n","      <td>0.000297</td>\n","      <td>-0.001284</td>\n","      <td>0.000485</td>\n","      <td>-0.001056</td>\n","      <td>-0.002393</td>\n","      <td>0.000868</td>\n","      <td>0.002491</td>\n","      <td>0.002799</td>\n","      <td>-0.000201</td>\n","      <td>-0.001307</td>\n","      <td>-0.000316</td>\n","      <td>-0.001018</td>\n","      <td>-0.001045</td>\n","      <td>0.000414</td>\n","      <td>-0.001369</td>\n","      <td>0.000409</td>\n","      <td>0.000602</td>\n","      <td>-0.003738</td>\n","      <td>0.001007</td>\n","      <td>0.001238</td>\n","      <td>-0.000555</td>\n","      <td>-0.000887</td>\n","      <td>0.000540</td>\n","      <td>0.000731</td>\n","      <td>0.001242</td>\n","      <td>-0.000098</td>\n","      <td>0.000208</td>\n","      <td>-7.826730e-04</td>\n","      <td>-0.000893</td>\n","      <td>0.000021</td>\n","      <td>0.000365</td>\n","      <td>-0.001477</td>\n","      <td>-0.001080</td>\n","      <td>0.002125</td>\n","      <td>-0.000431</td>\n","      <td>0.000626</td>\n","      <td>-0.000204</td>\n","      <td>0.001503</td>\n","      <td>-0.001048</td>\n","      <td>-0.000269</td>\n","      <td>-0.001128</td>\n","      <td>0.003119</td>\n","      <td>-0.000402</td>\n","      <td>0.000178</td>\n","      <td>-0.000781</td>\n","      <td>-0.000424</td>\n","      <td>0.000903</td>\n","      <td>-0.009369</td>\n","      <td>0.002768</td>\n","      <td>0.001170</td>\n","      <td>-0.001582</td>\n","      <td>0.001101</td>\n","      <td>-0.001706</td>\n","      <td>-0.000091</td>\n","      <td>-0.005400</td>\n","      <td>0.000601</td>\n","      <td>0.001082</td>\n","      <td>-0.000233</td>\n","      <td>-0.000045</td>\n","      <td>0.000268</td>\n","      <td>0.000824</td>\n","      <td>0.002958</td>\n","      <td>-0.000028</td>\n","      <td>-0.001380</td>\n","      <td>0.000567</td>\n","      <td>0.000152</td>\n","      <td>0.001126</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>-0.000367</td>\n","      <td>-0.000057</td>\n","      <td>0.000553</td>\n","      <td>-0.001005</td>\n","      <td>0.000073</td>\n","      <td>0.000991</td>\n","      <td>-0.001433</td>\n","      <td>0.001639</td>\n","      <td>0.001208</td>\n","      <td>0.000703</td>\n","      <td>-0.000135</td>\n","      <td>-0.001660</td>\n","      <td>0.000116</td>\n","      <td>-0.002004</td>\n","      <td>-0.001612</td>\n","      <td>0.000541</td>\n","      <td>-0.001707</td>\n","      <td>0.000487</td>\n","      <td>0.000504</td>\n","      <td>-0.005261</td>\n","      <td>-0.000538</td>\n","      <td>-0.000736</td>\n","      <td>0.075924</td>\n","      <td>-0.001389</td>\n","      <td>0.000826</td>\n","      <td>0.000243</td>\n","      <td>0.002401</td>\n","      <td>-0.001659</td>\n","      <td>0.000713</td>\n","      <td>6.167032e-04</td>\n","      <td>0.000752</td>\n","      <td>-0.001155</td>\n","      <td>-0.000002</td>\n","      <td>0.004459</td>\n","      <td>0.000130</td>\n","      <td>0.000337</td>\n","      <td>-0.002536</td>\n","      <td>0.000980</td>\n","      <td>0.000395</td>\n","      <td>0.002309</td>\n","      <td>0.000665</td>\n","      <td>-0.000415</td>\n","      <td>0.001801</td>\n","      <td>-0.001838</td>\n","      <td>0.001137</td>\n","      <td>-0.001298</td>\n","      <td>0.000560</td>\n","      <td>0.000301</td>\n","      <td>0.004132</td>\n","      <td>-0.001449</td>\n","      <td>0.004042</td>\n","      <td>0.001479</td>\n","      <td>0.001012</td>\n","      <td>0.001598</td>\n","      <td>0.000869</td>\n","      <td>-0.000581</td>\n","      <td>-0.000266</td>\n","      <td>0.000547</td>\n","      <td>-0.000713</td>\n","      <td>-0.001450</td>\n","      <td>0.000534</td>\n","      <td>-0.001127</td>\n","      <td>0.000161</td>\n","      <td>0.001155</td>\n","      <td>-0.002232</td>\n","      <td>0.000911</td>\n","      <td>0.000297</td>\n","      <td>-0.000051</td>\n","      <td>0.000858</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>-0.001298</td>\n","      <td>-0.000306</td>\n","      <td>0.001095</td>\n","      <td>0.002734</td>\n","      <td>0.002897</td>\n","      <td>0.000695</td>\n","      <td>0.000393</td>\n","      <td>0.000887</td>\n","      <td>0.001147</td>\n","      <td>-0.012540</td>\n","      <td>-0.000021</td>\n","      <td>-0.001430</td>\n","      <td>-0.000510</td>\n","      <td>-0.000621</td>\n","      <td>0.000285</td>\n","      <td>-0.000185</td>\n","      <td>0.000181</td>\n","      <td>0.001456</td>\n","      <td>0.000511</td>\n","      <td>-0.000245</td>\n","      <td>-0.002103</td>\n","      <td>0.000155</td>\n","      <td>-0.000607</td>\n","      <td>0.000594</td>\n","      <td>-0.000725</td>\n","      <td>0.001189</td>\n","      <td>-0.000495</td>\n","      <td>0.002572</td>\n","      <td>-0.001726</td>\n","      <td>-3.081860e-04</td>\n","      <td>-0.000333</td>\n","      <td>-0.006171</td>\n","      <td>-0.001059</td>\n","      <td>-0.000346</td>\n","      <td>0.000330</td>\n","      <td>-0.000009</td>\n","      <td>0.000265</td>\n","      <td>0.017157</td>\n","      <td>-0.001426</td>\n","      <td>0.001596</td>\n","      <td>0.000099</td>\n","      <td>0.000812</td>\n","      <td>-0.000440</td>\n","      <td>0.000766</td>\n","      <td>-0.000879</td>\n","      <td>0.000472</td>\n","      <td>-0.000785</td>\n","      <td>0.000384</td>\n","      <td>-0.000470</td>\n","      <td>0.001915</td>\n","      <td>-0.000462</td>\n","      <td>0.000266</td>\n","      <td>-0.000540</td>\n","      <td>0.001648</td>\n","      <td>-0.002368</td>\n","      <td>-0.000366</td>\n","      <td>0.000295</td>\n","      <td>0.000084</td>\n","      <td>0.002685</td>\n","      <td>-0.000056</td>\n","      <td>-0.000840</td>\n","      <td>-0.000028</td>\n","      <td>-0.001829</td>\n","      <td>-0.001120</td>\n","      <td>-0.000028</td>\n","      <td>-0.000102</td>\n","      <td>-0.000156</td>\n","      <td>-0.001136</td>\n","      <td>-0.000301</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>-0.001386</td>\n","      <td>0.001515</td>\n","      <td>0.000894</td>\n","      <td>-0.005358</td>\n","      <td>0.001380</td>\n","      <td>-0.003374</td>\n","      <td>0.000736</td>\n","      <td>0.000216</td>\n","      <td>-0.000597</td>\n","      <td>-0.000423</td>\n","      <td>-0.000651</td>\n","      <td>-0.000536</td>\n","      <td>-0.000161</td>\n","      <td>-0.001321</td>\n","      <td>0.001589</td>\n","      <td>-0.000149</td>\n","      <td>-0.001435</td>\n","      <td>0.001342</td>\n","      <td>0.000117</td>\n","      <td>-0.001416</td>\n","      <td>0.000060</td>\n","      <td>0.000423</td>\n","      <td>-0.000669</td>\n","      <td>0.001797</td>\n","      <td>0.001038</td>\n","      <td>-0.002386</td>\n","      <td>-0.000665</td>\n","      <td>-0.000528</td>\n","      <td>0.003252</td>\n","      <td>-4.347329e-03</td>\n","      <td>-0.000827</td>\n","      <td>0.001078</td>\n","      <td>-0.000259</td>\n","      <td>-0.001719</td>\n","      <td>-0.000071</td>\n","      <td>0.004244</td>\n","      <td>0.001308</td>\n","      <td>0.001823</td>\n","      <td>-0.012164</td>\n","      <td>-0.001202</td>\n","      <td>-0.003348</td>\n","      <td>0.000820</td>\n","      <td>-0.000765</td>\n","      <td>0.003476</td>\n","      <td>-0.001067</td>\n","      <td>0.001569</td>\n","      <td>0.000088</td>\n","      <td>0.000382</td>\n","      <td>-0.001031</td>\n","      <td>0.001000</td>\n","      <td>0.001576</td>\n","      <td>-0.000116</td>\n","      <td>-0.001858</td>\n","      <td>0.000256</td>\n","      <td>0.003232</td>\n","      <td>0.002149</td>\n","      <td>0.000045</td>\n","      <td>0.000143</td>\n","      <td>0.002952</td>\n","      <td>-0.000700</td>\n","      <td>0.000884</td>\n","      <td>0.001939</td>\n","      <td>-0.001356</td>\n","      <td>-0.000084</td>\n","      <td>0.005645</td>\n","      <td>-0.001302</td>\n","      <td>0.000820</td>\n","      <td>-0.000087</td>\n","      <td>0.000681</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>-0.001673</td>\n","      <td>-0.000238</td>\n","      <td>0.000555</td>\n","      <td>0.002977</td>\n","      <td>0.001828</td>\n","      <td>0.000354</td>\n","      <td>-0.001360</td>\n","      <td>0.000785</td>\n","      <td>0.001986</td>\n","      <td>-0.001801</td>\n","      <td>-0.000341</td>\n","      <td>-0.000607</td>\n","      <td>0.000099</td>\n","      <td>-0.001869</td>\n","      <td>-0.001686</td>\n","      <td>-0.000395</td>\n","      <td>0.000436</td>\n","      <td>0.000170</td>\n","      <td>0.000466</td>\n","      <td>-0.002103</td>\n","      <td>0.000688</td>\n","      <td>0.000049</td>\n","      <td>-0.000372</td>\n","      <td>0.002414</td>\n","      <td>-0.000413</td>\n","      <td>0.003520</td>\n","      <td>-0.000206</td>\n","      <td>0.001091</td>\n","      <td>-0.000430</td>\n","      <td>-7.888302e-07</td>\n","      <td>-0.000268</td>\n","      <td>-0.001050</td>\n","      <td>0.000765</td>\n","      <td>0.003270</td>\n","      <td>-0.000723</td>\n","      <td>0.000027</td>\n","      <td>-0.001562</td>\n","      <td>-0.001560</td>\n","      <td>0.005138</td>\n","      <td>0.000600</td>\n","      <td>0.000717</td>\n","      <td>-0.000325</td>\n","      <td>0.000365</td>\n","      <td>-0.001132</td>\n","      <td>-0.000106</td>\n","      <td>-0.000297</td>\n","      <td>0.001900</td>\n","      <td>0.000214</td>\n","      <td>0.002072</td>\n","      <td>-0.006242</td>\n","      <td>0.000179</td>\n","      <td>-0.000650</td>\n","      <td>-0.000841</td>\n","      <td>0.000400</td>\n","      <td>-0.000399</td>\n","      <td>-0.005746</td>\n","      <td>-0.001826</td>\n","      <td>-0.000022</td>\n","      <td>-0.002100</td>\n","      <td>0.000150</td>\n","      <td>-0.000776</td>\n","      <td>-0.001190</td>\n","      <td>-0.000267</td>\n","      <td>0.004009</td>\n","      <td>0.001184</td>\n","      <td>0.000042</td>\n","      <td>0.001739</td>\n","      <td>-0.001084</td>\n","      <td>-0.000172</td>\n","    </tr>\n","    <tr>\n","      <th>...</th>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","      <td>...</td>\n","    </tr>\n","    <tr>\n","      <th>12813</th>\n","      <td>-0.001386</td>\n","      <td>-0.000267</td>\n","      <td>0.000137</td>\n","      <td>-0.001674</td>\n","      <td>0.000790</td>\n","      <td>-0.001308</td>\n","      <td>0.000424</td>\n","      <td>0.001035</td>\n","      <td>0.002967</td>\n","      <td>0.000528</td>\n","      <td>-0.000286</td>\n","      <td>-0.000184</td>\n","      <td>-0.000475</td>\n","      <td>-0.002127</td>\n","      <td>-0.003310</td>\n","      <td>0.000022</td>\n","      <td>-0.000705</td>\n","      <td>0.000028</td>\n","      <td>0.000566</td>\n","      <td>-0.009401</td>\n","      <td>0.000040</td>\n","      <td>0.000133</td>\n","      <td>0.000358</td>\n","      <td>-0.000377</td>\n","      <td>0.000545</td>\n","      <td>-0.000629</td>\n","      <td>0.008500</td>\n","      <td>0.000536</td>\n","      <td>-0.000387</td>\n","      <td>-8.383766e-06</td>\n","      <td>0.001061</td>\n","      <td>0.002014</td>\n","      <td>-0.000614</td>\n","      <td>0.000729</td>\n","      <td>0.000339</td>\n","      <td>0.000908</td>\n","      <td>-0.003020</td>\n","      <td>-0.000684</td>\n","      <td>-0.001757</td>\n","      <td>0.002168</td>\n","      <td>-0.000157</td>\n","      <td>0.001572</td>\n","      <td>0.001676</td>\n","      <td>0.001422</td>\n","      <td>-0.000334</td>\n","      <td>0.000242</td>\n","      <td>-0.002029</td>\n","      <td>0.000091</td>\n","      <td>0.008237</td>\n","      <td>0.001407</td>\n","      <td>0.003542</td>\n","      <td>0.003316</td>\n","      <td>-0.000325</td>\n","      <td>0.002573</td>\n","      <td>-0.000085</td>\n","      <td>-0.002265</td>\n","      <td>0.001824</td>\n","      <td>0.000290</td>\n","      <td>0.000980</td>\n","      <td>-0.000399</td>\n","      <td>-0.000073</td>\n","      <td>0.000149</td>\n","      <td>-0.000030</td>\n","      <td>-0.000357</td>\n","      <td>-0.005483</td>\n","      <td>-0.000141</td>\n","      <td>-0.000050</td>\n","      <td>0.000281</td>\n","      <td>0.000085</td>\n","    </tr>\n","    <tr>\n","      <th>12814</th>\n","      <td>-0.000879</td>\n","      <td>0.000928</td>\n","      <td>-0.000324</td>\n","      <td>-0.000008</td>\n","      <td>-0.002774</td>\n","      <td>-0.001143</td>\n","      <td>0.002148</td>\n","      <td>0.000363</td>\n","      <td>-0.007701</td>\n","      <td>0.001909</td>\n","      <td>-0.000595</td>\n","      <td>-0.000112</td>\n","      <td>-0.000325</td>\n","      <td>-0.001438</td>\n","      <td>0.002683</td>\n","      <td>0.000275</td>\n","      <td>-0.001136</td>\n","      <td>-0.000290</td>\n","      <td>-0.000249</td>\n","      <td>0.001186</td>\n","      <td>0.001028</td>\n","      <td>0.001402</td>\n","      <td>-0.000376</td>\n","      <td>-0.000437</td>\n","      <td>-0.000128</td>\n","      <td>-0.001547</td>\n","      <td>-0.000893</td>\n","      <td>0.004491</td>\n","      <td>-0.000240</td>\n","      <td>-1.664882e-03</td>\n","      <td>0.000048</td>\n","      <td>0.002891</td>\n","      <td>0.000589</td>\n","      <td>-0.002657</td>\n","      <td>-0.001386</td>\n","      <td>0.001325</td>\n","      <td>0.002121</td>\n","      <td>0.001198</td>\n","      <td>-0.006189</td>\n","      <td>0.000433</td>\n","      <td>0.001269</td>\n","      <td>0.000857</td>\n","      <td>-0.000291</td>\n","      <td>0.000702</td>\n","      <td>0.000734</td>\n","      <td>0.001030</td>\n","      <td>0.000593</td>\n","      <td>0.000798</td>\n","      <td>-0.001239</td>\n","      <td>-0.000521</td>\n","      <td>0.003076</td>\n","      <td>-0.002144</td>\n","      <td>0.000486</td>\n","      <td>0.001234</td>\n","      <td>0.000053</td>\n","      <td>0.000593</td>\n","      <td>-0.000347</td>\n","      <td>-0.000598</td>\n","      <td>0.000471</td>\n","      <td>-0.000696</td>\n","      <td>0.001251</td>\n","      <td>-0.000367</td>\n","      <td>-0.001500</td>\n","      <td>0.000108</td>\n","      <td>0.006508</td>\n","      <td>-0.003649</td>\n","      <td>0.000566</td>\n","      <td>-0.000006</td>\n","      <td>-0.001073</td>\n","    </tr>\n","    <tr>\n","      <th>12815</th>\n","      <td>0.000157</td>\n","      <td>0.002493</td>\n","      <td>-0.000976</td>\n","      <td>0.004423</td>\n","      <td>0.002160</td>\n","      <td>-0.000876</td>\n","      <td>0.006307</td>\n","      <td>0.000164</td>\n","      <td>-0.002306</td>\n","      <td>-0.001813</td>\n","      <td>-0.000802</td>\n","      <td>-0.002237</td>\n","      <td>-0.000122</td>\n","      <td>-0.000805</td>\n","      <td>0.000320</td>\n","      <td>-0.000296</td>\n","      <td>-0.000887</td>\n","      <td>0.000792</td>\n","      <td>-0.000269</td>\n","      <td>0.002234</td>\n","      <td>0.001510</td>\n","      <td>0.002313</td>\n","      <td>-0.000354</td>\n","      <td>0.000225</td>\n","      <td>0.000375</td>\n","      <td>-0.002125</td>\n","      <td>-0.001877</td>\n","      <td>-0.001228</td>\n","      <td>-0.003278</td>\n","      <td>-1.001015e-03</td>\n","      <td>-0.000804</td>\n","      <td>0.000509</td>\n","      <td>0.009263</td>\n","      <td>0.000284</td>\n","      <td>-0.001416</td>\n","      <td>-0.002303</td>\n","      <td>0.000790</td>\n","      <td>-0.000548</td>\n","      <td>-0.007247</td>\n","      <td>-0.002367</td>\n","      <td>0.001609</td>\n","      <td>-0.001397</td>\n","      <td>0.000638</td>\n","      <td>0.004603</td>\n","      <td>0.000264</td>\n","      <td>0.002149</td>\n","      <td>-0.002899</td>\n","      <td>0.000641</td>\n","      <td>0.001796</td>\n","      <td>0.001736</td>\n","      <td>-0.001408</td>\n","      <td>-0.000321</td>\n","      <td>-0.001484</td>\n","      <td>0.001990</td>\n","      <td>-0.002550</td>\n","      <td>0.000515</td>\n","      <td>-0.000559</td>\n","      <td>-0.000009</td>\n","      <td>-0.000484</td>\n","      <td>0.000667</td>\n","      <td>0.000272</td>\n","      <td>-0.003701</td>\n","      <td>-0.000552</td>\n","      <td>0.005719</td>\n","      <td>0.004666</td>\n","      <td>0.003933</td>\n","      <td>0.001102</td>\n","      <td>0.000492</td>\n","      <td>-0.001795</td>\n","    </tr>\n","    <tr>\n","      <th>12816</th>\n","      <td>-0.000967</td>\n","      <td>0.000242</td>\n","      <td>-0.000617</td>\n","      <td>-0.002130</td>\n","      <td>-0.001000</td>\n","      <td>-0.002394</td>\n","      <td>0.001332</td>\n","      <td>0.001266</td>\n","      <td>0.003549</td>\n","      <td>0.003340</td>\n","      <td>-0.000537</td>\n","      <td>-0.001071</td>\n","      <td>-0.000673</td>\n","      <td>-0.001399</td>\n","      <td>-0.003830</td>\n","      <td>-0.000179</td>\n","      <td>0.000446</td>\n","      <td>0.000467</td>\n","      <td>0.000286</td>\n","      <td>-0.008399</td>\n","      <td>-0.002361</td>\n","      <td>0.000918</td>\n","      <td>0.000253</td>\n","      <td>-0.001874</td>\n","      <td>-0.000104</td>\n","      <td>-0.000586</td>\n","      <td>0.007003</td>\n","      <td>-0.000465</td>\n","      <td>-0.000478</td>\n","      <td>3.236285e-04</td>\n","      <td>0.001797</td>\n","      <td>0.001056</td>\n","      <td>0.000134</td>\n","      <td>0.004071</td>\n","      <td>0.000429</td>\n","      <td>0.002253</td>\n","      <td>-0.003139</td>\n","      <td>-0.001379</td>\n","      <td>0.006497</td>\n","      <td>0.000424</td>\n","      <td>0.000464</td>\n","      <td>0.000002</td>\n","      <td>0.001922</td>\n","      <td>0.001866</td>\n","      <td>0.000963</td>\n","      <td>-0.000706</td>\n","      <td>-0.001462</td>\n","      <td>0.000319</td>\n","      <td>0.009540</td>\n","      <td>0.000335</td>\n","      <td>0.003920</td>\n","      <td>0.005964</td>\n","      <td>0.000530</td>\n","      <td>0.002777</td>\n","      <td>0.002351</td>\n","      <td>-0.001881</td>\n","      <td>0.001857</td>\n","      <td>0.000187</td>\n","      <td>-0.000744</td>\n","      <td>-0.001658</td>\n","      <td>0.000308</td>\n","      <td>-0.001079</td>\n","      <td>0.001018</td>\n","      <td>0.002235</td>\n","      <td>-0.002130</td>\n","      <td>-0.001773</td>\n","      <td>-0.000062</td>\n","      <td>-0.001300</td>\n","      <td>0.001802</td>\n","    </tr>\n","    <tr>\n","      <th>12817</th>\n","      <td>-0.002101</td>\n","      <td>0.001221</td>\n","      <td>0.000603</td>\n","      <td>-0.002366</td>\n","      <td>0.003143</td>\n","      <td>-0.001599</td>\n","      <td>0.003491</td>\n","      <td>0.000486</td>\n","      <td>-0.002523</td>\n","      <td>0.003308</td>\n","      <td>-0.000351</td>\n","      <td>-0.003635</td>\n","      <td>-0.000216</td>\n","      <td>-0.002195</td>\n","      <td>-0.002047</td>\n","      <td>-0.000440</td>\n","      <td>0.001976</td>\n","      <td>0.001360</td>\n","      <td>0.000005</td>\n","      <td>-0.011312</td>\n","      <td>0.000739</td>\n","      <td>-0.000880</td>\n","      <td>-0.000283</td>\n","      <td>-0.001371</td>\n","      <td>-0.000685</td>\n","      <td>-0.000182</td>\n","      <td>-0.001500</td>\n","      <td>0.000619</td>\n","      <td>0.000247</td>\n","      <td>-1.254558e-03</td>\n","      <td>0.000711</td>\n","      <td>0.002036</td>\n","      <td>0.001327</td>\n","      <td>-0.000213</td>\n","      <td>0.001164</td>\n","      <td>0.001108</td>\n","      <td>-0.000713</td>\n","      <td>0.002031</td>\n","      <td>-0.002670</td>\n","      <td>0.002876</td>\n","      <td>0.000604</td>\n","      <td>-0.012780</td>\n","      <td>0.000319</td>\n","      <td>0.001192</td>\n","      <td>0.000009</td>\n","      <td>0.001497</td>\n","      <td>-0.000087</td>\n","      <td>0.000128</td>\n","      <td>-0.001542</td>\n","      <td>-0.001659</td>\n","      <td>0.002596</td>\n","      <td>-0.005955</td>\n","      <td>-0.001380</td>\n","      <td>0.002031</td>\n","      <td>-0.001480</td>\n","      <td>0.000278</td>\n","      <td>-0.001813</td>\n","      <td>0.000175</td>\n","      <td>-0.000203</td>\n","      <td>0.002372</td>\n","      <td>0.001198</td>\n","      <td>-0.001801</td>\n","      <td>0.000011</td>\n","      <td>-0.000270</td>\n","      <td>-0.002237</td>\n","      <td>0.001159</td>\n","      <td>0.001297</td>\n","      <td>-0.000661</td>\n","      <td>-0.000068</td>\n","    </tr>\n","  </tbody>\n","</table>\n","<p>12818 rows × 69 columns</p>\n","</div>"],"text/plain":["             0         1         2   ...        66        67        68\n","0     -0.001331  0.001008  0.000297  ...  0.000567  0.000152  0.001126\n","1     -0.000367 -0.000057  0.000553  ...  0.000297 -0.000051  0.000858\n","2     -0.001298 -0.000306  0.001095  ... -0.000156 -0.001136 -0.000301\n","3     -0.001386  0.001515  0.000894  ...  0.000820 -0.000087  0.000681\n","4     -0.001673 -0.000238  0.000555  ...  0.001739 -0.001084 -0.000172\n","...         ...       ...       ...  ...       ...       ...       ...\n","12813 -0.001386 -0.000267  0.000137  ... -0.000050  0.000281  0.000085\n","12814 -0.000879  0.000928 -0.000324  ...  0.000566 -0.000006 -0.001073\n","12815  0.000157  0.002493 -0.000976  ...  0.001102  0.000492 -0.001795\n","12816 -0.000967  0.000242 -0.000617  ... -0.000062 -0.001300  0.001802\n","12817 -0.002101  0.001221  0.000603  ...  0.001297 -0.000661 -0.000068\n","\n","[12818 rows x 69 columns]"]},"metadata":{"tags":[]},"execution_count":32}]},{"cell_type":"markdown","metadata":{"id":"J3BPuKGEfOOV"},"source":["## Encode the data sets"]},{"cell_type":"code","metadata":{"id":"S0-KdvUyb1a7"},"source":["X_train_encoded_features = encoder_model.predict(X_train_scaled)\n","X_valid_encoded_features = encoder_model.predict(X_valid_scaled)\n","X_test_encoded_features = encoder_model.predict(X_test_scaled)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"JW86J_uSfUsx"},"source":["## Classifier Model"]},{"cell_type":"code","metadata":{"id":"8petY812cMI0","colab":{"base_uri":"https://localhost:8080/","height":323},"executionInfo":{"status":"ok","timestamp":1603800961576,"user_tz":240,"elapsed":1652,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"11c49594-c6dc-4c27-ceac-3abedf14607f"},"source":["model = Sequential()\n","model.add(Input(shape=(X_train_encoded_features.shape[1], \n","                       X_train_encoded_features.shape[2]), \n","                name='input'))\n","model.add(Conv1D(filters=16, \n","                 kernel_size=4,\n","                 activation='relu', \n","                 padding='valid'))\n","model.add(MaxPool1D(pool_size=4, \n","                    padding='valid'))\n","model.add(Flatten())\n","model.add(Dense(units=16, \n","                activation='relu'))\n","model.add(Dense(units=1, \n","                activation='sigmoid', \n","                name='output'))\n","model.summary()"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Model: \"sequential_2\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","conv1d_2 (Conv1D)            (None, 17, 16)            4432      \n","_________________________________________________________________\n","max_pooling1d_2 (MaxPooling1 (None, 4, 16)             0         \n","_________________________________________________________________\n","flatten_2 (Flatten)          (None, 64)                0         \n","_________________________________________________________________\n","dense_3 (Dense)              (None, 16)                1040      \n","_________________________________________________________________\n","output (Dense)               (None, 1)                 17        \n","=================================================================\n","Total params: 5,489\n","Trainable params: 5,489\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"G2Ls5OYrcWF7","colab":{"base_uri":"https://localhost:8080/","height":1000},"executionInfo":{"status":"ok","timestamp":1603801080742,"user_tz":240,"elapsed":118150,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"7485a7e0-cb0b-43a9-de82-d169fda88354"},"source":["model.compile(optimizer='adam',\n","              loss='binary_crossentropy',\n","              metrics=[\n","                  'accuracy',\n","                  tf.keras.metrics.Recall(),\n","                  pm.F1Score(),\n","                  pm.FalsePositiveRate()\n","              ])\n","history = model.fit(x=X_train_encoded_features,\n","                    y=y_train,\n","                    batch_size=128,\n","                    epochs=150,\n","                    validation_data=(X_valid_encoded_features, y_valid),\n","                    verbose=1).history"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Epoch 1/150\n","102/102 [==============================] - 1s 10ms/step - loss: 0.3818 - accuracy: 0.9808 - recall_2: 0.0057 - f1_score: 0.0080 - false_positive_rate: 0.0059 - val_loss: 0.0892 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 2/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0748 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0625 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 3/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0706 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0619 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 4/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0697 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0614 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 5/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0691 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0609 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 6/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0682 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0605 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 7/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0674 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 8/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0667 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0594 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 9/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0661 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0592 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 10/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0656 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0591 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 11/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0655 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0592 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 12/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0652 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0590 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 13/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0651 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0591 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 14/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0651 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0590 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 15/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0651 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0590 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 16/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0648 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0592 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 17/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0644 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0591 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 18/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0644 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0590 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 19/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0645 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0594 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 20/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0643 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0592 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 21/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0644 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0594 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 22/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0643 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0594 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 23/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0641 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0593 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 24/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0641 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 25/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0642 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0592 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 26/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0639 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0609 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 27/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0640 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0593 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 28/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0639 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0593 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 29/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0638 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0600 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 30/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0640 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0591 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 31/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0638 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0591 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 32/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0638 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0598 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 33/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0640 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0598 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 34/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0638 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0600 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 35/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0637 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0598 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 36/150\n","102/102 [==============================] - 1s 9ms/step - loss: 0.0637 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0592 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 37/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0635 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0599 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 38/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0634 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0612 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 39/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0639 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0603 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 40/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0636 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0593 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 41/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0637 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0593 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 42/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0633 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0594 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 43/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0638 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 44/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0633 - accuracy: 0.9867 - recall_2: 0.0057 - f1_score: 0.0114 - false_positive_rate: 0.0000e+00 - val_loss: 0.0593 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 45/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0628 - accuracy: 0.9868 - recall_2: 0.0115 - f1_score: 0.0227 - false_positive_rate: 0.0000e+00 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 46/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0632 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0595 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 47/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0632 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0593 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 48/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0633 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0595 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 49/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0631 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0594 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 50/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0632 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0600 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 51/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0631 - accuracy: 0.9867 - recall_2: 0.0057 - f1_score: 0.0114 - false_positive_rate: 0.0000e+00 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 52/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0632 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 53/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0631 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0601 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 54/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0629 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0596 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 55/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0629 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0594 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 56/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0627 - accuracy: 0.9866 - recall_2: 0.0000e+00 - f1_score: 0.0000e+00 - false_positive_rate: 0.0000e+00 - val_loss: 0.0613 - val_accuracy: 0.9880 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 6.2266e-04\n","Epoch 57/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0627 - accuracy: 0.9867 - recall_2: 0.0057 - f1_score: 0.0114 - false_positive_rate: 0.0000e+00 - val_loss: 0.0597 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 58/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0625 - accuracy: 0.9868 - recall_2: 0.0115 - f1_score: 0.0227 - false_positive_rate: 0.0000e+00 - val_loss: 0.0596 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 59/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0625 - accuracy: 0.9868 - recall_2: 0.0115 - f1_score: 0.0227 - false_positive_rate: 0.0000e+00 - val_loss: 0.0603 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 60/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0626 - accuracy: 0.9867 - recall_2: 0.0057 - f1_score: 0.0114 - false_positive_rate: 0.0000e+00 - val_loss: 0.0598 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 61/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0622 - accuracy: 0.9868 - recall_2: 0.0115 - f1_score: 0.0227 - false_positive_rate: 0.0000e+00 - val_loss: 0.0600 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 62/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0623 - accuracy: 0.9867 - recall_2: 0.0057 - f1_score: 0.0114 - false_positive_rate: 0.0000e+00 - val_loss: 0.0599 - val_accuracy: 0.9880 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 6.2266e-04\n","Epoch 63/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0622 - accuracy: 0.9868 - recall_2: 0.0115 - f1_score: 0.0227 - false_positive_rate: 0.0000e+00 - val_loss: 0.0602 - val_accuracy: 0.9880 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 6.2266e-04\n","Epoch 64/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0620 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0339 - false_positive_rate: 0.0000e+00 - val_loss: 0.0595 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 65/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0621 - accuracy: 0.9867 - recall_2: 0.0115 - f1_score: 0.0226 - false_positive_rate: 7.8015e-05 - val_loss: 0.0599 - val_accuracy: 0.9880 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 6.2266e-04\n","Epoch 66/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0620 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0339 - false_positive_rate: 0.0000e+00 - val_loss: 0.0595 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 67/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0618 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0337 - false_positive_rate: 7.8015e-05 - val_loss: 0.0605 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 68/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0616 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0337 - false_positive_rate: 7.8015e-05 - val_loss: 0.0595 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 69/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0617 - accuracy: 0.9868 - recall_2: 0.0230 - f1_score: 0.0447 - false_positive_rate: 7.8015e-05 - val_loss: 0.0594 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 70/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0614 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0339 - false_positive_rate: 0.0000e+00 - val_loss: 0.0599 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 71/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0613 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0337 - false_positive_rate: 7.8015e-05 - val_loss: 0.0596 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 72/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0612 - accuracy: 0.9867 - recall_2: 0.0115 - f1_score: 0.0226 - false_positive_rate: 7.8015e-05 - val_loss: 0.0595 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 73/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0610 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0339 - false_positive_rate: 0.0000e+00 - val_loss: 0.0599 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 74/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0610 - accuracy: 0.9868 - recall_2: 0.0230 - f1_score: 0.0447 - false_positive_rate: 7.8015e-05 - val_loss: 0.0595 - val_accuracy: 0.9886 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 0.0000e+00\n","Epoch 75/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0610 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0339 - false_positive_rate: 0.0000e+00 - val_loss: 0.0598 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 76/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0609 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0595 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 77/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0608 - accuracy: 0.9868 - recall_2: 0.0172 - f1_score: 0.0337 - false_positive_rate: 7.8015e-05 - val_loss: 0.0597 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 78/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0606 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0600 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 79/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0606 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0608 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 80/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0604 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0595 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 81/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0603 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0601 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 82/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0602 - accuracy: 0.9868 - recall_2: 0.0230 - f1_score: 0.0447 - false_positive_rate: 7.8015e-05 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 83/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0603 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 84/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0602 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0605 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 85/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0604 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0608 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 86/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0600 - accuracy: 0.9870 - recall_2: 0.0345 - f1_score: 0.0663 - false_positive_rate: 7.8015e-05 - val_loss: 0.0599 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 87/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0598 - accuracy: 0.9868 - recall_2: 0.0230 - f1_score: 0.0447 - false_positive_rate: 7.8015e-05 - val_loss: 0.0605 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 88/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0596 - accuracy: 0.9870 - recall_2: 0.0345 - f1_score: 0.0663 - false_positive_rate: 7.8015e-05 - val_loss: 0.0598 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 89/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0598 - accuracy: 0.9870 - recall_2: 0.0345 - f1_score: 0.0663 - false_positive_rate: 7.8015e-05 - val_loss: 0.0600 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 90/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0597 - accuracy: 0.9870 - recall_2: 0.0345 - f1_score: 0.0663 - false_positive_rate: 7.8015e-05 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 91/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0596 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0596 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 92/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0593 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0595 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 93/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0591 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0597 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 94/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0592 - accuracy: 0.9870 - recall_2: 0.0345 - f1_score: 0.0663 - false_positive_rate: 7.8015e-05 - val_loss: 0.0595 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 95/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0592 - accuracy: 0.9871 - recall_2: 0.0402 - f1_score: 0.0769 - false_positive_rate: 7.8015e-05 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 96/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0590 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0598 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 97/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0590 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0599 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 98/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0590 - accuracy: 0.9871 - recall_2: 0.0402 - f1_score: 0.0769 - false_positive_rate: 7.8015e-05 - val_loss: 0.0615 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 99/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0591 - accuracy: 0.9869 - recall_2: 0.0287 - f1_score: 0.0556 - false_positive_rate: 7.8015e-05 - val_loss: 0.0639 - val_accuracy: 0.9880 - val_recall_2: 0.0270 - val_f1_score: 0.0488 - val_false_positive_rate: 9.3400e-04\n","Epoch 100/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0588 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0870 - false_positive_rate: 1.5603e-04 - val_loss: 0.0608 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 101/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0596 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0598 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 102/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0587 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0600 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 103/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0585 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0598 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 104/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0586 - accuracy: 0.9870 - recall_2: 0.0345 - f1_score: 0.0663 - false_positive_rate: 7.8015e-05 - val_loss: 0.0611 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 105/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0586 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0607 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 106/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0583 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0597 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 107/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0581 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0598 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 108/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0581 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0605 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 109/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0582 - accuracy: 0.9871 - recall_2: 0.0402 - f1_score: 0.0769 - false_positive_rate: 7.8015e-05 - val_loss: 0.0603 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 110/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0580 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0599 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 111/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0579 - accuracy: 0.9870 - recall_2: 0.0345 - f1_score: 0.0663 - false_positive_rate: 7.8015e-05 - val_loss: 0.0598 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 112/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0576 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0605 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 113/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0578 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0597 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 114/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0577 - accuracy: 0.9871 - recall_2: 0.0402 - f1_score: 0.0769 - false_positive_rate: 7.8015e-05 - val_loss: 0.0600 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 115/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0575 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0598 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 116/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0574 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0601 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 117/150\n","102/102 [==============================] - 1s 9ms/step - loss: 0.0572 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0604 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 118/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0572 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0610 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 119/150\n","102/102 [==============================] - 1s 9ms/step - loss: 0.0572 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0607 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 120/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0575 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0611 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 121/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0573 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0601 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 122/150\n","102/102 [==============================] - 1s 9ms/step - loss: 0.0570 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0602 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 123/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0567 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0607 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 124/150\n","102/102 [==============================] - 1s 9ms/step - loss: 0.0573 - accuracy: 0.9870 - recall_2: 0.0460 - f1_score: 0.0865 - false_positive_rate: 2.3405e-04 - val_loss: 0.0605 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 125/150\n","102/102 [==============================] - 1s 9ms/step - loss: 0.0568 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0602 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 126/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0565 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0607 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 127/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0566 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0612 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 128/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0564 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0601 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 129/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0563 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0636 - val_accuracy: 0.9877 - val_recall_2: 0.0270 - val_f1_score: 0.0476 - val_false_positive_rate: 0.0012\n","Epoch 130/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0567 - accuracy: 0.9871 - recall_2: 0.0402 - f1_score: 0.0769 - false_positive_rate: 7.8015e-05 - val_loss: 0.0613 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 131/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0562 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0606 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 132/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0564 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0609 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 133/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0564 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0609 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 134/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0565 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0613 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 135/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0563 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0623 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 136/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0564 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0603 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 137/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0558 - accuracy: 0.9871 - recall_2: 0.0517 - f1_score: 0.0973 - false_positive_rate: 1.5603e-04 - val_loss: 0.0606 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 138/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0563 - accuracy: 0.9870 - recall_2: 0.0402 - f1_score: 0.0765 - false_positive_rate: 1.5603e-04 - val_loss: 0.0608 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 139/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0563 - accuracy: 0.9872 - recall_2: 0.0575 - f1_score: 0.1075 - false_positive_rate: 1.5603e-04 - val_loss: 0.0604 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 140/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0558 - accuracy: 0.9871 - recall_2: 0.0517 - f1_score: 0.0973 - false_positive_rate: 1.5603e-04 - val_loss: 0.0608 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 141/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0558 - accuracy: 0.9871 - recall_2: 0.0460 - f1_score: 0.0874 - false_positive_rate: 7.8015e-05 - val_loss: 0.0604 - val_accuracy: 0.9883 - val_recall_2: 0.0000e+00 - val_f1_score: 0.0000e+00 - val_false_positive_rate: 3.1133e-04\n","Epoch 142/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0557 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0606 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 143/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0557 - accuracy: 0.9872 - recall_2: 0.0517 - f1_score: 0.0978 - false_positive_rate: 7.8015e-05 - val_loss: 0.0612 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 144/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0554 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0604 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 145/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0558 - accuracy: 0.9874 - recall_2: 0.0632 - f1_score: 0.1183 - false_positive_rate: 7.8015e-05 - val_loss: 0.0620 - val_accuracy: 0.9883 - val_recall_2: 0.0270 - val_f1_score: 0.0500 - val_false_positive_rate: 6.2266e-04\n","Epoch 146/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0553 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0606 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 147/150\n","102/102 [==============================] - 1s 7ms/step - loss: 0.0553 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0610 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 148/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0548 - accuracy: 0.9874 - recall_2: 0.0632 - f1_score: 0.1183 - false_positive_rate: 7.8015e-05 - val_loss: 0.0617 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 149/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0554 - accuracy: 0.9873 - recall_2: 0.0575 - f1_score: 0.1081 - false_positive_rate: 7.8015e-05 - val_loss: 0.0613 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n","Epoch 150/150\n","102/102 [==============================] - 1s 8ms/step - loss: 0.0549 - accuracy: 0.9871 - recall_2: 0.0575 - f1_score: 0.1070 - false_positive_rate: 2.3405e-04 - val_loss: 0.0614 - val_accuracy: 0.9886 - val_recall_2: 0.0270 - val_f1_score: 0.0513 - val_false_positive_rate: 3.1133e-04\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"Vvm_JOLTfaFR"},"source":["## Results"]},{"cell_type":"code","metadata":{"id":"4ilBw19icZfb","colab":{"base_uri":"https://localhost:8080/","height":295},"executionInfo":{"status":"ok","timestamp":1603789426973,"user_tz":240,"elapsed":499197,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"705e0287-4978-4e0a-c0a3-62d8f2de1165"},"source":["plt, fig = sp.plot_metric(history, metric='loss')\n","fig.savefig('sparse_cnn_encoder_classifier_loss.pdf', bbox_inches='tight')"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"id":"W0i-OLibc4V_","colab":{"base_uri":"https://localhost:8080/","height":297},"executionInfo":{"status":"ok","timestamp":1603792379650,"user_tz":240,"elapsed":748,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"31277191-f9ef-4e6b-b441-5183d8c40d31"},"source":["plt, fig = sp.plot_metric(history, metric='f1_score', ylim=[0., 1.])\n","# fig.savefig('sparse_cnn_encoder_classifier_f1_score.pdf', bbox_inches='tight')"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"id":"iryDXBZPc7TO","colab":{"base_uri":"https://localhost:8080/","height":294},"executionInfo":{"status":"ok","timestamp":1603789427455,"user_tz":240,"elapsed":499651,"user":{"displayName":"Chitta Ranjan","photoUrl":"","userId":"03244148267376825421"}},"outputId":"c5caa63f-813e-449b-d5f4-c3eac0d831b6"},"source":["plt, fig = sp.plot_model_recall_fpr(history)\n","# fig.savefig('sparse_cnn_encoder_classifier_recall_fpr.pdf', bbox_inches='tight')"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]}]}