Experiment_predict_performance_from_input_RDM.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Experiment: predict performance from input RDM.ipynb",
      "version": "0.3.2",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d-TYfeIDlyF8",
        "colab_type": "text"
      },
      "source": [
        "#Classifying with RDM matrix for 800 models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bS-9c5ZwlG8o",
        "colab_type": "text"
      },
      "source": [
        "The question: how much information do input RDMs contain? Is this information enough to train a simple classifier to predict a model performance from its input RDM?\n",
        "\n",
        "The answer: yes"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-aK7grMKoNx6",
        "colab_type": "code",
        "outputId": "4be47301-d338-4850-b966-b527f14f51dc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from google.colab import drive \n",
        "drive.mount('/content/gdrive', force_remount=True)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IJXQBs46hbTY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import urllib.request\n",
        "import pickle\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from matplotlib.patches import Rectangle\n",
        "import seaborn\n",
        "import pandas as pd\n",
        "import sklearn\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.ensemble import RandomForestRegressor\n",
        "\n",
        "ROOT_PATH = 'gdrive/My Drive/NI Project - RSA/'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mxdPoZZriRcV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "with open (ROOT_PATH + 'rdm800.pkl', 'rb') as fp:\n",
        "    rdm = pickle.load(fp)\n",
        "    \n",
        "with open (ROOT_PATH + 'acc_list.pkl', 'rb') as fp:\n",
        "    acc_list = pickle.load(fp)\n",
        "    \n",
        "with open (ROOT_PATH + 'corr_distances_dict800.pkl', 'rb') as fp:\n",
        "    corr_distances_dict = pickle.load(fp)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KMLi3vV9kw1u",
        "colab_type": "text"
      },
      "source": [
        "In the RDM models are sorted in accordance with increasing accuracy"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g_zhzefGiZ9R",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "acc_only = np.array(np.array(acc_list)[:, 0], dtype='float32')\n",
        "files_only = np.array(np.array(acc_list)[:, 1])\n",
        "\n",
        "acc_sorted_idx = np.argsort(acc_only)\n",
        "acc_sorted = acc_only[acc_sorted_idx]\n",
        "files_sorted = files_only[acc_sorted_idx]\n",
        "n = len(acc_sorted)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8HuJ-CZlpRO-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "rdms = np.ones((n, 19900)) * -17 # num of models * linearised upper triangle of input rdm\n",
        "for i in range(n):\n",
        "  cur_corr_distances = corr_distances_dict[files_sorted[i]]\n",
        "  cur_input_rdm = cur_corr_distances[np.triu_indices(cur_corr_distances.shape[0], k = 1)]\n",
        "  rdms[i] = cur_input_rdm\n",
        "  \n",
        "seed = 42\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4cpqOZ3rurVl",
        "colab_type": "text"
      },
      "source": [
        "##Classification"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P0d1uC9elhiy",
        "colab_type": "text"
      },
      "source": [
        "Classification doesn't work well, but we do through away information when constructing classes."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lw8XojgJqRZo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "labels = (acc_sorted * 10)//1\n",
        "X_train, X_test, y_train, y_test = train_test_split(rdms, labels, test_size=0.2, random_state=seed, shuffle=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GRJ1KF5ntdfO",
        "colab_type": "code",
        "outputId": "69bb0379-4dc0-42ed-ff49-4664943754b0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        }
      },
      "source": [
        "print(X_train.shape)\n",
        "print(y_train.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(636, 19900)\n",
            "(636,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BIY6qbsKsAvL",
        "colab_type": "code",
        "outputId": "8f0f12ec-b620-4e04-93a3-9e3afa50cc0b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 134
        }
      },
      "source": [
        "clf = RandomForestClassifier(n_estimators=100, max_depth=2, random_state=seed)\n",
        "clf.fit(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
              "                       max_depth=2, max_features='auto', max_leaf_nodes=None,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=1, min_samples_split=2,\n",
              "                       min_weight_fraction_leaf=0.0, n_estimators=100,\n",
              "                       n_jobs=None, oob_score=False, random_state=42, verbose=0,\n",
              "                       warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fJZGaa2isMuW",
        "colab_type": "code",
        "outputId": "fa8960f7-62a6-4276-b715-7ca508ab14df",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        }
      },
      "source": [
        "print('Train')\n",
        "print(clf.score(X_train, y_train))\n",
        "print('Test')\n",
        "print(clf.score(X_test, y_test))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train\n",
            "0.5283018867924528\n",
            "Test\n",
            "0.5283018867924528\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KHWdfAf4slPu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for i in range(n):\n",
        "#   print(rdms[i].shape)\n",
        "#   print(np.expand_dims(rdms[i], axis=0).shape)\n",
        "  print(clf.predict(np.expand_dims(rdms[i], axis=0)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EZD6DC6bt1oJ",
        "colab_type": "text"
      },
      "source": [
        "##Regression"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ifboBVCgmCpQ",
        "colab_type": "text"
      },
      "source": [
        "Regression works well"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EciI7o0lv17t",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn import linear_model\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from sklearn.neural_network import MLPRegressor"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m5Ya0SCLuxhS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(rdms, acc_sorted, test_size=0.2, random_state=seed, shuffle=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7z5mlAhNu0gY",
        "colab_type": "code",
        "outputId": "848c7634-ca0e-4b3d-e1eb-bcc94fb61486",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        }
      },
      "source": [
        "print(X_train.shape)\n",
        "print(y_train.shape)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(636, 19900)\n",
            "(636,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jYW7oQjOu08l",
        "colab_type": "code",
        "outputId": "8771052c-2d85-4e73-8d02-9b37f8ec8d02",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        }
      },
      "source": [
        "# regr = linear_model.LinearRegression(normalize=True)\n",
        "# regr = MLPRegressor(hidden_layer_sizes=(50, 50, 20), verbose=True)\n",
        "regr = linear_model.BayesianRidge(normalize=True)\n",
        "# regr = linear_model.Ridge(alpha=15, normalize=True)\n",
        "regr.fit(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "BayesianRidge(alpha_1=1e-06, alpha_2=1e-06, compute_score=False, copy_X=True,\n",
              "              fit_intercept=True, lambda_1=1e-06, lambda_2=1e-06, n_iter=300,\n",
              "              normalize=True, tol=0.001, verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qJm_cIKRvQA1",
        "colab_type": "code",
        "outputId": "1b43567e-ce91-482e-9601-ffbf426866f2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 101
        }
      },
      "source": [
        "print('R2 coefficients')\n",
        "print('Train')\n",
        "print(regr.score(X_train, y_train))\n",
        "print('Test')\n",
        "print(regr.score(X_test, y_test))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "R2 coefficients\n",
            "Train\n",
            "0.8599034314970877\n",
            "Test\n",
            "0.7928269787488849\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UtrSyGg5wcz3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for i in range(len(X_test)):\n",
        "  if y_test[i] < 0.7:\n",
        "    continue\n",
        "#   print(X_test[i].shape)\n",
        "  cur_predict = regr.predict(np.expand_dims(X_test[i], axis=0))[0]\n",
        "  print(cur_predict, y_test[i], np.abs(cur_predict - y_test[i]), sep='\\t\\t')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iBylhFKP6b12",
        "colab_type": "code",
        "outputId": "a1ce3392-1abe-460e-ef55-faf8381324c8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        }
      },
      "source": [
        "print('Train')\n",
        "y_pred = regr.predict(X_train)\n",
        "print(np.sqrt(mean_squared_error(y_train * 100, y_pred * 100)))\n",
        "print((np.abs(y_train * 100 - y_pred * 100)).mean())"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train\n",
            "6.351936001823895\n",
            "4.984817203482253\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Z_GT-Gg0nQA",
        "colab_type": "code",
        "outputId": "31211f7e-7820-4506-88a9-3711000ff200",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        }
      },
      "source": [
        "print('Test')\n",
        "y_pred = regr.predict(X_test)\n",
        "print(np.sqrt(mean_squared_error(y_test * 100, y_pred * 100)))\n",
        "print((np.abs(y_test * 100 - y_pred * 100)).mean())"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test\n",
            "7.912808320028136\n",
            "6.053572574480843\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j58W6xW1mSZr",
        "colab_type": "text"
      },
      "source": [
        "Compare true values with predictions"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MWUyo03hv7_E",
        "colab_type": "code",
        "outputId": "01d68035-b32e-4de9-c0ec-0e4a9bbb64b3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 13373
        }
      },
      "source": [
        "for i in range(n):\n",
        "  print(regr.predict(np.expand_dims(rdms[i], axis=0)), acc_sorted[i])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0.30464563] 0.1998\n",
            "[0.30464515] 0.1998\n",
            "[0.30462512] 0.2014\n",
            "[0.32312711] 0.28959998\n",
            "[0.30533898] 0.2898\n",
            "[0.31503603] 0.2908\n",
            "[0.42711813] 0.2911\n",
            "[0.33702136] 0.29569998\n",
            "[0.35842253] 0.29569998\n",
            "[0.4919167] 0.2958\n",
            "[0.39445566] 0.296\n",
            "[0.50614992] 0.296\n",
            "[0.35082256] 0.2961\n",
            "[0.34894511] 0.2963\n",
            "[0.34559634] 0.2963\n",
            "[0.34381126] 0.2963\n",
            "[0.34694609] 0.2963\n",
            "[0.49820307] 0.2963\n",
            "[0.37728284] 0.29639998\n",
            "[0.37904013] 0.29639998\n",
            "[0.34752656] 0.29639998\n",
            "[0.34612525] 0.29639998\n",
            "[0.36796572] 0.2965\n",
            "[0.35730828] 0.29659998\n",
            "[0.34634308] 0.29659998\n",
            "[0.36855134] 0.2967\n",
            "[0.36753733] 0.2967\n",
            "[0.40829032] 0.2968\n",
            "[0.47384599] 0.2968\n",
            "[0.39016155] 0.2969\n",
            "[0.36131639] 0.297\n",
            "[0.30700505] 0.297\n",
            "[0.3969904] 0.297\n",
            "[0.34209156] 0.29709998\n",
            "[0.34968411] 0.29709998\n",
            "[0.34707982] 0.29709998\n",
            "[0.30474121] 0.29709998\n",
            "[0.38555104] 0.2972\n",
            "[0.36304739] 0.29729998\n",
            "[0.34118101] 0.29749998\n",
            "[0.37305061] 0.2976\n",
            "[0.37018057] 0.2979\n",
            "[0.38212433] 0.2981\n",
            "[0.3757373] 0.2988\n",
            "[0.44742384] 0.3125\n",
            "[0.3406505] 0.32189998\n",
            "[0.35423941] 0.3453\n",
            "[0.54156937] 0.3588\n",
            "[0.41375728] 0.3635\n",
            "[0.4608578] 0.3674\n",
            "[0.39110273] 0.3715\n",
            "[0.4384578] 0.37269998\n",
            "[0.59492057] 0.3739\n",
            "[0.46986331] 0.3741\n",
            "[0.3513535] 0.3744\n",
            "[0.38940511] 0.3748\n",
            "[0.46802017] 0.37539998\n",
            "[0.4841847] 0.3755\n",
            "[0.40900816] 0.376\n",
            "[0.62224576] 0.37649998\n",
            "[0.48923583] 0.3768\n",
            "[0.54845874] 0.3771\n",
            "[0.53100552] 0.3782\n",
            "[0.39304232] 0.37829998\n",
            "[0.49726288] 0.37849998\n",
            "[0.38126255] 0.37939999\n",
            "[0.42451538] 0.3796\n",
            "[0.56757553] 0.3796\n",
            "[0.51641604] 0.37989998\n",
            "[0.42004778] 0.37989998\n",
            "[0.43349667] 0.3804\n",
            "[0.38290943] 0.3809\n",
            "[0.45901345] 0.38099998\n",
            "[0.4521726] 0.3813\n",
            "[0.48297638] 0.38169998\n",
            "[0.36440634] 0.3818\n",
            "[0.51027244] 0.3818\n",
            "[0.35425233] 0.38189998\n",
            "[0.35470759] 0.382\n",
            "[0.58066068] 0.382\n",
            "[0.47585375] 0.3822\n",
            "[0.54932369] 0.3823\n",
            "[0.37298037] 0.38239998\n",
            "[0.43101327] 0.3827\n",
            "[0.3511564] 0.38279998\n",
            "[0.40786511] 0.38349998\n",
            "[0.35831486] 0.3839\n",
            "[0.4230596] 0.3841\n",
            "[0.4702542] 0.3849\n",
            "[0.4838103] 0.3858\n",
            "[0.52914192] 0.3859\n",
            "[0.45408616] 0.3859\n",
            "[0.39891059] 0.38599998\n",
            "[0.45074073] 0.3861\n",
            "[0.3967763] 0.3861\n",
            "[0.3881199] 0.3865\n",
            "[0.56765644] 0.3865\n",
            "[0.56981409] 0.387\n",
            "[0.39746428] 0.387\n",
            "[0.41067952] 0.38709998\n",
            "[0.4213329] 0.3872\n",
            "[0.43706674] 0.3874\n",
            "[0.41990312] 0.3874\n",
            "[0.35810038] 0.3875\n",
            "[0.44007884] 0.3876\n",
            "[0.45627733] 0.3877\n",
            "[0.37503276] 0.38779998\n",
            "[0.40015463] 0.3881\n",
            "[0.46216394] 0.3883\n",
            "[0.5188208] 0.3883\n",
            "[0.52587035] 0.3884\n",
            "[0.44180689] 0.38889998\n",
            "[0.39589451] 0.389\n",
            "[0.44542778] 0.3891\n",
            "[0.48948179] 0.3892\n",
            "[0.3835231] 0.3892\n",
            "[0.40231003] 0.3893\n",
            "[0.45642146] 0.3893\n",
            "[0.47963338] 0.38939998\n",
            "[0.43124858] 0.3897\n",
            "[0.40788458] 0.3897\n",
            "[0.39691777] 0.3897\n",
            "[0.38445595] 0.3897\n",
            "[0.4711658] 0.38979998\n",
            "[0.46105969] 0.3899\n",
            "[0.43553408] 0.39\n",
            "[0.43223872] 0.39\n",
            "[0.39600713] 0.3902\n",
            "[0.45867493] 0.3902\n",
            "[0.47624487] 0.39029998\n",
            "[0.47127217] 0.39029998\n",
            "[0.42550297] 0.3904\n",
            "[0.44934931] 0.39049998\n",
            "[0.42127442] 0.39069998\n",
            "[0.41244368] 0.3908\n",
            "[0.48814658] 0.3908\n",
            "[0.40968702] 0.3908\n",
            "[0.4821584] 0.3909\n",
            "[0.40543679] 0.39159998\n",
            "[0.45453963] 0.39159998\n",
            "[0.38594334] 0.39159998\n",
            "[0.37798124] 0.3917\n",
            "[0.47865832] 0.3919\n",
            "[0.46657336] 0.39249998\n",
            "[0.47191174] 0.3926\n",
            "[0.39687961] 0.3927\n",
            "[0.38227475] 0.3928\n",
            "[0.39455039] 0.3928\n",
            "[0.46462937] 0.39339998\n",
            "[0.4528507] 0.3949\n",
            "[0.43478667] 0.3958\n",
            "[0.48575176] 0.4037\n",
            "[0.49283944] 0.42819998\n",
            "[0.36542956] 0.4477\n",
            "[0.4072577] 0.4479\n",
            "[0.61979529] 0.4493\n",
            "[0.45752435] 0.4527\n",
            "[0.48899485] 0.4556\n",
            "[0.53166089] 0.4558\n",
            "[0.44240778] 0.4567\n",
            "[0.54583729] 0.45799997\n",
            "[0.49967268] 0.45839998\n",
            "[0.46760869] 0.45999998\n",
            "[0.60444008] 0.4606\n",
            "[0.57312003] 0.46379998\n",
            "[0.51282787] 0.46429998\n",
            "[0.46501969] 0.4664\n",
            "[0.47786908] 0.4669\n",
            "[0.61205579] 0.4669\n",
            "[0.48235872] 0.46699998\n",
            "[0.38623153] 0.4671\n",
            "[0.59159048] 0.46789998\n",
            "[0.55489191] 0.46859998\n",
            "[0.51838442] 0.46859998\n",
            "[0.53952922] 0.46879998\n",
            "[0.5034284] 0.46899998\n",
            "[0.42369224] 0.46899998\n",
            "[0.56471767] 0.4692\n",
            "[0.54569583] 0.4702\n",
            "[0.67097894] 0.4707\n",
            "[0.44281631] 0.47079998\n",
            "[0.49569031] 0.4716\n",
            "[0.66760373] 0.47169998\n",
            "[0.58203817] 0.4721\n",
            "[0.55638663] 0.4725\n",
            "[0.47342983] 0.4727\n",
            "[0.52467879] 0.473\n",
            "[0.57541307] 0.47329998\n",
            "[0.54524314] 0.4734\n",
            "[0.51294588] 0.4748\n",
            "[0.61888648] 0.4754\n",
            "[0.73441807] 0.4759\n",
            "[0.48495734] 0.4761\n",
            "[0.44781475] 0.4766\n",
            "[0.4366283] 0.4768\n",
            "[0.46145261] 0.47689998\n",
            "[0.53100777] 0.477\n",
            "[0.51166327] 0.47709998\n",
            "[0.50506884] 0.4773\n",
            "[0.6082558] 0.478\n",
            "[0.42516128] 0.4781\n",
            "[0.49326753] 0.4781\n",
            "[0.6313345] 0.47829998\n",
            "[0.45565393] 0.47869998\n",
            "[0.44550101] 0.4788\n",
            "[0.56042765] 0.4789\n",
            "[0.46901693] 0.479\n",
            "[0.55962254] 0.47919998\n",
            "[0.48790073] 0.4793\n",
            "[0.59323231] 0.47939998\n",
            "[0.50328607] 0.4795\n",
            "[0.4832404] 0.48029998\n",
            "[0.45395636] 0.48029998\n",
            "[0.47766769] 0.48029998\n",
            "[0.46487143] 0.4804\n",
            "[0.44335967] 0.48049998\n",
            "[0.46735953] 0.48099998\n",
            "[0.51067866] 0.48099998\n",
            "[0.49298529] 0.4815\n",
            "[0.53931158] 0.4816\n",
            "[0.47889601] 0.48189998\n",
            "[0.4239262] 0.48209998\n",
            "[0.54381621] 0.48319998\n",
            "[0.54282792] 0.48319998\n",
            "[0.49747614] 0.4834\n",
            "[0.39112416] 0.4843\n",
            "[0.4402507] 0.4847\n",
            "[0.5260714] 0.48479998\n",
            "[0.44000781] 0.4849\n",
            "[0.50048483] 0.4861\n",
            "[0.42712048] 0.48639998\n",
            "[0.39127032] 0.4869\n",
            "[0.52695951] 0.4877\n",
            "[0.37544762] 0.4935\n",
            "[0.48474955] 0.4965\n",
            "[0.56185651] 0.50909996\n",
            "[0.56084062] 0.51629996\n",
            "[0.53105089] 0.5205\n",
            "[0.49945363] 0.5261\n",
            "[0.56502193] 0.5335\n",
            "[0.46246832] 0.5351\n",
            "[0.56933695] 0.536\n",
            "[0.49599713] 0.54109997\n",
            "[0.58180751] 0.5437\n",
            "[0.6214478] 0.5473\n",
            "[0.4242991] 0.55039996\n",
            "[0.51923162] 0.55079997\n",
            "[0.4776413] 0.5512\n",
            "[0.53273925] 0.55149996\n",
            "[0.45035479] 0.5516\n",
            "[0.88813264] 0.5517\n",
            "[0.44428754] 0.55289996\n",
            "[0.45687883] 0.555\n",
            "[0.56829335] 0.55509996\n",
            "[0.55069707] 0.5553\n",
            "[0.63300194] 0.5555\n",
            "[0.60364103] 0.556\n",
            "[0.58225161] 0.5561\n",
            "[0.45929155] 0.5566\n",
            "[0.59824813] 0.5571\n",
            "[0.47170701] 0.55799997\n",
            "[0.57853635] 0.55799997\n",
            "[0.48847763] 0.55799997\n",
            "[0.54814122] 0.5581\n",
            "[0.6103167] 0.5589\n",
            "[0.46491449] 0.5592\n",
            "[0.45670626] 0.5593\n",
            "[0.51541133] 0.5597\n",
            "[0.54211216] 0.55979997\n",
            "[0.54926344] 0.5604\n",
            "[0.56829523] 0.5615\n",
            "[0.65167792] 0.56159997\n",
            "[0.60599198] 0.56159997\n",
            "[0.49271303] 0.5617\n",
            "[0.61446661] 0.5618\n",
            "[0.50094505] 0.5621\n",
            "[0.48404795] 0.5625\n",
            "[0.59073213] 0.5628\n",
            "[0.62609447] 0.5629\n",
            "[0.6049972] 0.5643\n",
            "[0.56110148] 0.5645\n",
            "[0.65050501] 0.5645\n",
            "[0.56283003] 0.5645\n",
            "[0.54751177] 0.5656\n",
            "[0.55900547] 0.5656\n",
            "[0.54935871] 0.5657\n",
            "[0.43555894] 0.5658\n",
            "[0.72449183] 0.566\n",
            "[0.55959881] 0.5669\n",
            "[0.56210133] 0.5671\n",
            "[0.53917844] 0.56729996\n",
            "[0.54434456] 0.5679\n",
            "[0.50557665] 0.56799996\n",
            "[0.68467138] 0.5681\n",
            "[0.51257739] 0.5682\n",
            "[0.44022102] 0.5694\n",
            "[0.48641338] 0.5697\n",
            "[0.54988588] 0.57\n",
            "[0.47186262] 0.5701\n",
            "[0.58855613] 0.5701\n",
            "[0.52560767] 0.5703\n",
            "[0.45214724] 0.5707\n",
            "[0.58654946] 0.5715\n",
            "[0.54616601] 0.57199997\n",
            "[0.47894254] 0.5733\n",
            "[0.60636082] 0.5736\n",
            "[0.62561174] 0.57409996\n",
            "[0.53432496] 0.5749\n",
            "[0.53494791] 0.57519996\n",
            "[0.5430575] 0.5753\n",
            "[0.52929529] 0.57589996\n",
            "[0.54926562] 0.576\n",
            "[0.58440654] 0.5765\n",
            "[0.56352328] 0.5767\n",
            "[0.59156676] 0.5782\n",
            "[0.52114814] 0.5782\n",
            "[0.61246562] 0.58919996\n",
            "[0.60771811] 0.5999\n",
            "[0.54572382] 0.606\n",
            "[0.67978439] 0.6125\n",
            "[0.55001725] 0.61439997\n",
            "[0.63604447] 0.6145\n",
            "[0.52815934] 0.6153\n",
            "[0.58427768] 0.6177\n",
            "[0.5936638] 0.62149996\n",
            "[0.67820318] 0.62299997\n",
            "[0.56004572] 0.62329996\n",
            "[0.55329524] 0.62619996\n",
            "[0.58632843] 0.62689996\n",
            "[0.55307793] 0.6274\n",
            "[0.56240004] 0.6281\n",
            "[0.7281224] 0.6296\n",
            "[0.5058775] 0.6303\n",
            "[0.69192749] 0.6306\n",
            "[0.57870689] 0.6313\n",
            "[0.59811866] 0.6317\n",
            "[0.57643449] 0.63229996\n",
            "[0.50028089] 0.6326\n",
            "[0.61160764] 0.63299996\n",
            "[0.64828767] 0.6331\n",
            "[0.63598531] 0.6332\n",
            "[0.52141693] 0.6337\n",
            "[0.64489485] 0.634\n",
            "[0.72402126] 0.6344\n",
            "[0.64780653] 0.63449997\n",
            "[0.77416153] 0.63519996\n",
            "[0.59020463] 0.63549995\n",
            "[0.69733207] 0.63549995\n",
            "[0.73475225] 0.6361\n",
            "[0.58429395] 0.63699996\n",
            "[0.61856172] 0.63699996\n",
            "[0.65542893] 0.6383\n",
            "[0.54590853] 0.63909996\n",
            "[0.55704088] 0.63989997\n",
            "[0.61217261] 0.6401\n",
            "[0.65881977] 0.64059997\n",
            "[0.71264684] 0.64199996\n",
            "[0.66426701] 0.6423\n",
            "[0.55805352] 0.6432\n",
            "[0.70228842] 0.6434\n",
            "[0.59746195] 0.6434\n",
            "[0.61746605] 0.6435\n",
            "[0.65631463] 0.6451\n",
            "[0.63129625] 0.6451\n",
            "[0.59631294] 0.64519995\n",
            "[0.64144381] 0.6454\n",
            "[0.53981946] 0.64559996\n",
            "[0.67311071] 0.6468\n",
            "[0.65333598] 0.6472\n",
            "[0.66373706] 0.6475\n",
            "[0.74452403] 0.6476\n",
            "[0.56889281] 0.64779997\n",
            "[0.63644208] 0.64779997\n",
            "[0.56417279] 0.6487\n",
            "[0.60384871] 0.64879996\n",
            "[0.68340548] 0.64919996\n",
            "[0.55804744] 0.6504\n",
            "[0.63760031] 0.6505\n",
            "[0.57597602] 0.6505\n",
            "[0.61642865] 0.65099996\n",
            "[0.6431803] 0.6511\n",
            "[0.67289329] 0.65169996\n",
            "[0.71730141] 0.6522\n",
            "[0.63468695] 0.65349996\n",
            "[0.63872163] 0.6544\n",
            "[0.70164583] 0.6545\n",
            "[0.69450067] 0.655\n",
            "[0.63244131] 0.65529996\n",
            "[0.61302199] 0.65599996\n",
            "[0.75349776] 0.6561\n",
            "[0.74688647] 0.65669996\n",
            "[0.61079854] 0.6569\n",
            "[0.6208098] 0.6572\n",
            "[0.67506927] 0.6574\n",
            "[0.6426205] 0.6574\n",
            "[0.6200835] 0.65749997\n",
            "[0.64410004] 0.65819997\n",
            "[0.65159625] 0.6588\n",
            "[0.57866929] 0.6592\n",
            "[0.59596534] 0.6592\n",
            "[0.59644807] 0.6595\n",
            "[0.69805388] 0.6597\n",
            "[0.6364292] 0.6604\n",
            "[0.65486588] 0.6604\n",
            "[0.64908673] 0.66069996\n",
            "[0.69880757] 0.66069996\n",
            "[0.63066098] 0.66099995\n",
            "[0.63973885] 0.66099995\n",
            "[0.69136086] 0.6615\n",
            "[0.67204979] 0.6622\n",
            "[0.65028662] 0.6634\n",
            "[0.69941966] 0.66359997\n",
            "[0.66971718] 0.6644\n",
            "[0.63329587] 0.6649\n",
            "[0.67802135] 0.666\n",
            "[0.62381468] 0.6671\n",
            "[0.6867451] 0.6677\n",
            "[0.59292281] 0.66889995\n",
            "[0.61977528] 0.6702\n",
            "[0.71153291] 0.6721\n",
            "[0.5751815] 0.6778\n",
            "[0.77824666] 0.6881\n",
            "[0.64050716] 0.69049996\n",
            "[0.59355954] 0.691\n",
            "[0.60145795] 0.6996\n",
            "[0.68455542] 0.7026\n",
            "[0.73107253] 0.7035\n",
            "[0.70170555] 0.7037\n",
            "[0.70824119] 0.70629996\n",
            "[0.62172594] 0.70669997\n",
            "[0.71623902] 0.7073\n",
            "[0.63551189] 0.71099997\n",
            "[0.71525645] 0.7122\n",
            "[0.68814493] 0.7125\n",
            "[0.61202046] 0.71349996\n",
            "[0.72820215] 0.7154\n",
            "[0.68345127] 0.71559995\n",
            "[0.73827888] 0.71669996\n",
            "[0.69999658] 0.718\n",
            "[0.57347176] 0.71819997\n",
            "[0.72221109] 0.71819997\n",
            "[0.69446744] 0.71849996\n",
            "[0.62093687] 0.7197\n",
            "[0.63339885] 0.7201\n",
            "[0.72525158] 0.7204\n",
            "[0.74937144] 0.7209\n",
            "[0.7839475] 0.72139996\n",
            "[0.66919192] 0.7219\n",
            "[0.67408251] 0.722\n",
            "[0.7692601] 0.72209996\n",
            "[0.76976535] 0.72209996\n",
            "[0.58250651] 0.7223\n",
            "[0.65750457] 0.7224\n",
            "[0.75590017] 0.72249997\n",
            "[0.69743049] 0.7226\n",
            "[0.64810484] 0.72279996\n",
            "[0.82232735] 0.723\n",
            "[0.7126815] 0.7231\n",
            "[0.69925606] 0.72319996\n",
            "[0.73737842] 0.7234\n",
            "[0.69249884] 0.7242\n",
            "[0.78563074] 0.7249\n",
            "[0.69238522] 0.7251\n",
            "[0.66103612] 0.7255\n",
            "[0.76648378] 0.7263\n",
            "[0.69881433] 0.7266\n",
            "[0.67913717] 0.72679996\n",
            "[0.65853634] 0.72679996\n",
            "[0.80885018] 0.7269\n",
            "[0.76043166] 0.72709996\n",
            "[0.68868693] 0.72779995\n",
            "[0.73102787] 0.728\n",
            "[0.58733829] 0.7291\n",
            "[0.71599818] 0.7291\n",
            "[0.7010628] 0.7291\n",
            "[0.76636604] 0.72969997\n",
            "[0.74016317] 0.7308\n",
            "[0.714197] 0.7319\n",
            "[0.58343583] 0.7327\n",
            "[0.76941401] 0.73319995\n",
            "[0.66499949] 0.7337\n",
            "[0.62041101] 0.736\n",
            "[0.6429722] 0.7363\n",
            "[0.66517053] 0.7378\n",
            "[0.74677176] 0.7384\n",
            "[0.64494475] 0.7385\n",
            "[0.67568524] 0.7387\n",
            "[0.56823292] 0.74079996\n",
            "[0.75821332] 0.74079996\n",
            "[0.69421148] 0.7421\n",
            "[0.75723615] 0.7421\n",
            "[0.68206702] 0.7425\n",
            "[0.79724961] 0.7428\n",
            "[0.64577743] 0.7435\n",
            "[0.67119332] 0.74369997\n",
            "[0.61228452] 0.7438\n",
            "[0.68299907] 0.7439\n",
            "[0.69700667] 0.7441\n",
            "[0.66225455] 0.7442\n",
            "[0.74599113] 0.74439996\n",
            "[0.68342688] 0.74439996\n",
            "[0.71090366] 0.7448\n",
            "[0.70912663] 0.7448\n",
            "[0.64982475] 0.7456\n",
            "[0.82812375] 0.7456\n",
            "[0.67350475] 0.7457\n",
            "[0.76566492] 0.7457\n",
            "[0.72955512] 0.7461\n",
            "[0.71935033] 0.74619997\n",
            "[0.71504246] 0.7463\n",
            "[0.68485374] 0.74649996\n",
            "[0.64395956] 0.7467\n",
            "[0.64847383] 0.7468\n",
            "[0.77371166] 0.7468\n",
            "[0.72108069] 0.74689996\n",
            "[0.66617617] 0.74689996\n",
            "[0.70036093] 0.74689996\n",
            "[0.61618841] 0.747\n",
            "[0.77486252] 0.74719995\n",
            "[0.71521955] 0.74719995\n",
            "[0.75495098] 0.74729997\n",
            "[0.72029249] 0.74729997\n",
            "[0.61577549] 0.74729997\n",
            "[0.63315827] 0.7474\n",
            "[0.76025821] 0.7474\n",
            "[0.67312447] 0.7475\n",
            "[0.7621511] 0.7475\n",
            "[0.74630678] 0.74759996\n",
            "[0.72417609] 0.7478\n",
            "[0.76691981] 0.7478\n",
            "[0.75540101] 0.7479\n",
            "[0.77064543] 0.74799997\n",
            "[0.75246758] 0.7482\n",
            "[0.69261297] 0.7482\n",
            "[0.66291724] 0.7482\n",
            "[0.83250351] 0.74829996\n",
            "[0.68099838] 0.7484\n",
            "[0.76819529] 0.7484\n",
            "[0.78043075] 0.7485\n",
            "[0.70206159] 0.74869996\n",
            "[0.69755629] 0.74869996\n",
            "[0.59187865] 0.74869996\n",
            "[0.6929867] 0.7488\n",
            "[0.73607639] 0.74899995\n",
            "[0.74568897] 0.7492\n",
            "[0.70132226] 0.7492\n",
            "[0.68389176] 0.7493\n",
            "[0.67905591] 0.74939996\n",
            "[0.66947227] 0.74939996\n",
            "[0.71487579] 0.7495\n",
            "[0.70739059] 0.7497\n",
            "[0.75755263] 0.7497\n",
            "[0.7444728] 0.74979997\n",
            "[0.68222651] 0.74979997\n",
            "[0.71802164] 0.74979997\n",
            "[0.75269678] 0.7499\n",
            "[0.72576155] 0.7503\n",
            "[0.67492649] 0.7504\n",
            "[0.71664284] 0.75049996\n",
            "[0.77054418] 0.7507\n",
            "[0.68740476] 0.75079995\n",
            "[0.79895955] 0.7511\n",
            "[0.72459147] 0.75119996\n",
            "[0.72746554] 0.7513\n",
            "[0.66399272] 0.7515\n",
            "[0.69102655] 0.7517\n",
            "[0.77606802] 0.75229996\n",
            "[0.73709901] 0.7525\n",
            "[0.75625151] 0.75299996\n",
            "[0.73188439] 0.75799996\n",
            "[0.77637826] 0.7603\n",
            "[0.78711006] 0.7682\n",
            "[0.77359044] 0.77309996\n",
            "[0.61209555] 0.7739\n",
            "[0.74892904] 0.7743\n",
            "[0.74450828] 0.7804\n",
            "[0.74582371] 0.78209996\n",
            "[0.75653033] 0.7822\n",
            "[0.77402217] 0.7834\n",
            "[0.80895775] 0.7838\n",
            "[0.81265952] 0.7847\n",
            "[0.71485986] 0.7848\n",
            "[0.73451178] 0.7851\n",
            "[0.75908923] 0.7851\n",
            "[0.66621799] 0.7873\n",
            "[0.71212476] 0.7876\n",
            "[0.76179533] 0.7912\n",
            "[0.74036382] 0.79219997\n",
            "[0.87065457] 0.79289997\n",
            "[0.6014949] 0.7945\n",
            "[0.75527534] 0.7949\n",
            "[0.77017792] 0.7959\n",
            "[0.71703917] 0.7962\n",
            "[0.70543367] 0.79639995\n",
            "[0.77918456] 0.7966\n",
            "[0.71355228] 0.7974\n",
            "[0.78805598] 0.7978\n",
            "[0.66308491] 0.7978\n",
            "[0.79501076] 0.7992\n",
            "[0.71075583] 0.7994\n",
            "[0.7585299] 0.7996\n",
            "[0.76434059] 0.7998\n",
            "[0.71405064] 0.7999\n",
            "[0.74039165] 0.80079997\n",
            "[0.80710279] 0.8016\n",
            "[0.70267228] 0.8017\n",
            "[0.74641563] 0.8017\n",
            "[0.8917782] 0.80189997\n",
            "[0.61786507] 0.802\n",
            "[0.7443939] 0.80249995\n",
            "[0.7918438] 0.8027\n",
            "[0.70631901] 0.8032\n",
            "[0.82250406] 0.80329996\n",
            "[0.73241702] 0.80359995\n",
            "[0.75036888] 0.8038\n",
            "[0.81190479] 0.80399996\n",
            "[0.75569314] 0.80399996\n",
            "[0.70256397] 0.8041\n",
            "[0.73183744] 0.80469996\n",
            "[0.8289355] 0.8049\n",
            "[0.7760871] 0.805\n",
            "[0.77242131] 0.80509996\n",
            "[0.74949139] 0.8052\n",
            "[0.81493945] 0.80539995\n",
            "[0.69292307] 0.806\n",
            "[0.86369961] 0.80619997\n",
            "[0.70890479] 0.80619997\n",
            "[0.82243539] 0.80619997\n",
            "[0.74115893] 0.80649996\n",
            "[0.76969914] 0.80649996\n",
            "[0.75428386] 0.80649996\n",
            "[0.96480606] 0.8067\n",
            "[0.68410764] 0.80689996\n",
            "[0.7247009] 0.807\n",
            "[0.75359304] 0.8073\n",
            "[0.74841592] 0.8073\n",
            "[0.74778907] 0.8075\n",
            "[0.74957097] 0.8077\n",
            "[0.76807394] 0.8078\n",
            "[0.73646638] 0.8081\n",
            "[0.72666967] 0.8082\n",
            "[0.80566881] 0.80829996\n",
            "[0.71950298] 0.8085\n",
            "[0.77713623] 0.8085\n",
            "[0.78033912] 0.80869997\n",
            "[0.7346125] 0.80869997\n",
            "[0.69729213] 0.8089\n",
            "[0.77771084] 0.80899996\n",
            "[0.70676983] 0.8091\n",
            "[0.70759864] 0.8092\n",
            "[0.78489626] 0.8092\n",
            "[0.8003391] 0.8093\n",
            "[0.81695697] 0.8093\n",
            "[0.7762788] 0.8093\n",
            "[0.7080323] 0.80939996\n",
            "[0.76608945] 0.8095\n",
            "[0.74128395] 0.8096\n",
            "[0.70323237] 0.8096\n",
            "[0.77571717] 0.8096\n",
            "[0.6751176] 0.8096\n",
            "[0.87577247] 0.8096\n",
            "[0.7346808] 0.80979997\n",
            "[0.69913772] 0.81\n",
            "[0.79736103] 0.81009996\n",
            "[0.84054087] 0.81009996\n",
            "[0.77638681] 0.81009996\n",
            "[0.73836515] 0.81009996\n",
            "[0.73746577] 0.8102\n",
            "[0.77441286] 0.8102\n",
            "[0.77869078] 0.8102\n",
            "[0.89655539] 0.81049997\n",
            "[0.70718861] 0.81049997\n",
            "[0.73851867] 0.8106\n",
            "[0.8017067] 0.8106\n",
            "[0.65445048] 0.8107\n",
            "[0.75972488] 0.8107\n",
            "[0.78159714] 0.8107\n",
            "[0.72712238] 0.8109\n",
            "[0.76416562] 0.811\n",
            "[0.70029129] 0.8111\n",
            "[0.75936609] 0.8113\n",
            "[0.76473765] 0.8113\n",
            "[0.83130637] 0.8114\n",
            "[0.73218581] 0.8114\n",
            "[0.78009337] 0.81149995\n",
            "[0.80416497] 0.8118\n",
            "[0.84768415] 0.812\n",
            "[0.78448771] 0.8121\n",
            "[0.83676617] 0.8121\n",
            "[0.82547034] 0.8121\n",
            "[0.84473516] 0.81219995\n",
            "[0.77479458] 0.81229997\n",
            "[0.75610938] 0.81229997\n",
            "[0.87777965] 0.81229997\n",
            "[0.83168224] 0.8127\n",
            "[0.78911787] 0.8127\n",
            "[0.76857085] 0.8127\n",
            "[0.77171906] 0.8128\n",
            "[0.81198818] 0.8129\n",
            "[0.75452066] 0.8129\n",
            "[0.78918406] 0.81299996\n",
            "[0.74695001] 0.8132\n",
            "[0.82110972] 0.8132\n",
            "[0.80120615] 0.81329995\n",
            "[0.78530815] 0.81329995\n",
            "[0.7376879] 0.8134\n",
            "[0.81237861] 0.8134\n",
            "[0.78397795] 0.8134\n",
            "[0.89937108] 0.8135\n",
            "[0.77945415] 0.8135\n",
            "[0.77854974] 0.8135\n",
            "[0.88845618] 0.8136\n",
            "[0.80704186] 0.8136\n",
            "[0.78090777] 0.8138\n",
            "[0.74829246] 0.8139\n",
            "[0.87563382] 0.8139\n",
            "[0.79722491] 0.8139\n",
            "[0.8401456] 0.81399995\n",
            "[0.8703596] 0.8142\n",
            "[0.74829249] 0.8142\n",
            "[0.87147615] 0.8143\n",
            "[0.77229246] 0.8143\n",
            "[0.86723071] 0.8143\n",
            "[0.72216988] 0.8143\n",
            "[0.82031586] 0.8143\n",
            "[0.7936645] 0.81439996\n",
            "[0.88819144] 0.8145\n",
            "[0.93275696] 0.8145\n",
            "[0.84932271] 0.8146\n",
            "[0.69684294] 0.8147\n",
            "[0.74986777] 0.81479996\n",
            "[0.7528245] 0.81479996\n",
            "[0.79985782] 0.81479996\n",
            "[0.91339369] 0.81479996\n",
            "[0.79002127] 0.8149\n",
            "[0.79261943] 0.8152\n",
            "[0.80949089] 0.8153\n",
            "[0.88091814] 0.8153\n",
            "[0.87602178] 0.8154\n",
            "[0.78406786] 0.8154\n",
            "[0.75950045] 0.81549996\n",
            "[0.81774216] 0.81549996\n",
            "[0.68547668] 0.8157\n",
            "[0.90409887] 0.8157\n",
            "[0.8456147] 0.81579995\n",
            "[0.72527879] 0.81589997\n",
            "[0.77558278] 0.81589997\n",
            "[0.82662886] 0.816\n",
            "[0.82808777] 0.8161\n",
            "[0.74174386] 0.8161\n",
            "[0.84244413] 0.81619996\n",
            "[0.80278074] 0.8163\n",
            "[0.7087377] 0.8164\n",
            "[0.8144763] 0.8164\n",
            "[0.87125959] 0.8164\n",
            "[0.76742491] 0.8164\n",
            "[0.77693545] 0.8164\n",
            "[0.73923522] 0.81659997\n",
            "[0.84744907] 0.8167\n",
            "[0.80723811] 0.8167\n",
            "[0.80008642] 0.8167\n",
            "[0.73081851] 0.81689996\n",
            "[0.80514246] 0.81689996\n",
            "[0.76529548] 0.8171\n",
            "[0.78717641] 0.81729996\n",
            "[0.76830159] 0.8175\n",
            "[0.80467266] 0.8175\n",
            "[0.91092933] 0.8178\n",
            "[0.85568478] 0.8178\n",
            "[0.87376348] 0.8179\n",
            "[0.80856418] 0.8179\n",
            "[0.75899527] 0.8181\n",
            "[0.81625987] 0.8181\n",
            "[0.79546172] 0.8183\n",
            "[0.77505976] 0.8183\n",
            "[0.88752555] 0.81839997\n",
            "[0.81776541] 0.81839997\n",
            "[0.7942436] 0.8185\n",
            "[0.79564868] 0.8189\n",
            "[0.71501922] 0.819\n",
            "[0.83068357] 0.819\n",
            "[0.84124458] 0.8192\n",
            "[0.81905189] 0.8192\n",
            "[0.85229202] 0.8192\n",
            "[0.75424713] 0.81939995\n",
            "[0.74198387] 0.8197\n",
            "[0.8530947] 0.8197\n",
            "[0.79956547] 0.81979996\n",
            "[0.81700308] 0.8203\n",
            "[0.78015608] 0.8208\n",
            "[0.78030785] 0.8211\n",
            "[0.85306324] 0.8211\n",
            "[0.82452721] 0.82119995\n",
            "[0.89427298] 0.82199997\n",
            "[0.81216971] 0.8549\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cEqLPKVd9pwl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_pred = regr.predict(rdms)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Sv9N5dem9c2U",
        "colab_type": "code",
        "outputId": "9f44f458-6117-454d-a152-89f52f0a95c4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plt.title('Predicted vs true performance (models are sorted in increasing order)')\n",
        "\n",
        "plt.ylim(0, 1)\n",
        "plt.plot(y_pred, 'o')\n",
        "# plt.show()\n",
        "\n",
        "# plt.ylim(0, 1)\n",
        "plt.plot(acc_sorted, 'o')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}