RFC.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ce80c78c",
   "metadata": {},
   "source": [
    "We explored four different methods to train and test our datasets, each with its own set of advantages and challenges. The first approach involved Support Vector Machines (SVM) using both polynomial (poly) and Radial Basis Function (RBF) kernels. However, SVM proved to be time-consuming, especially with optimization attempts, resulting in a significant number of false negatives."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6b9f3a2",
   "metadata": {},
   "source": [
    "The second method employed AdaBoost, which exhibited a relatively faster training time ranging from 40 minutes to 2 hours. Despite its efficiency, adjusting parameters posed challenges."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "62563f54",
   "metadata": {},
   "source": [
    "Our third method utilized Catboost, which, according to our literature survey, was expected to outperform SVM and AdaBoost by at least five times in terms of speed. While it generally lived up to this expectation, there were instances where parameter tuning extended the training time to approximately an hour. The lowest false negative rate achieved was around 2K."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "986a05ca",
   "metadata": {},
   "source": [
    "The final method, Random Forest Classifier (RFC), emerged as the most efficient option for our dataset. RFC demonstrated exceptional speed, with parameter tuning resulting in minimal impact on training time. After fine-tuning, we achieved zero false negatives and a low incidence of false positives, completing dataset training in a maximum of 10 minutes per iteration."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f59c2895a580665",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "In the below cell we combine all the data from all the files. Each column in each file is stacked to the nm column to form a data of shape nx2\n",
    "Labels are added, all data is considered dead and labled 0, except for data from 2016skin.csv file which is live skin and labeled 1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "fbd90e5b903749c9",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-01-24T22:02:01.557300900Z",
     "start_time": "2024-01-24T22:01:33.691809800Z"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.ensemble import RandomForestClassifier as RFC\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix, precision_score, recall_score, ConfusionMatrixDisplay\n",
    "\n",
    "\n",
    "fL = './ReadShowLeather/data/Leder.csv'\n",
    "fH = './ReadShowLeather/data/Holz.csv'\n",
    "fF = './ReadShowLeather/data/Fleisch.csv'\n",
    "fS = '2016skin.csv'\n",
    "fM = '2016material.csv'\n",
    "fMF = '2016material-fake.csv'\n",
    "\n",
    "Files = [fL, fH, fF, fM, fMF, fS]\n",
    "decimals = [',',',',',','.','.','.']\n",
    "Y_values= [0,0,0,0,0,1]\n",
    "\n",
    "data = np.zeros((1,3))\n",
    "for i, f in enumerate(Files):\n",
    "    df = pd.read_csv(f, delimiter=';',decimal=decimals[i], index_col=0)\n",
    "    df = df.dropna()\n",
    "    for col in df.columns:\n",
    "        Y=np.ones((len(df[col]),1))*Y_values[i]\n",
    "        X1=df.index.to_numpy().reshape((-1,1))\n",
    "        X2=df[col].to_numpy().reshape((-1,1))\n",
    "        d=np.hstack((X1,X2,Y))\n",
    "        data = np.vstack((data,d))\n",
    "np.random.shuffle(data)\n",
    "data = np.delete(data, 0, 0)\n",
    "\n",
    "X= data[:,:2].reshape((-1,2))\n",
    "Y= data[:,2]\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "1ab86f90cc40665",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-01-24T22:02:36.692965100Z",
     "start_time": "2024-01-24T22:02:01.568999200Z"
    },
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 12 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "building tree 1 of 30\n",
      "building tree 2 of 30\n",
      "building tree 3 of 30\n",
      "building tree 4 of 30\n",
      "building tree 5 of 30\n",
      "building tree 6 of 30\n",
      "building tree 7 of 30\n",
      "building tree 8 of 30\n",
      "building tree 9 of 30\n",
      "building tree 10 of 30\n",
      "building tree 11 of 30\n",
      "building tree 12 of 30\n",
      "building tree 13 of 30\n",
      "building tree 14 of 30\n",
      "building tree 15 of 30\n",
      "building tree 16 of 30\n",
      "building tree 17 of 30\n",
      "building tree 18 of 30\n",
      "building tree 19 of 30building tree 20 of 30\n",
      "building tree 21 of 30\n",
      "building tree 22 of 30\n",
      "building tree 23 of 30\n",
      "building tree 24 of 30\n",
      "building tree 25 of 30\n",
      "building tree 26 of 30\n",
      "building tree 27 of 30\n",
      "building tree 28 of 30\n",
      "building tree 29 of 30\n",
      "building tree 30 of 30\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  23 out of  30 | elapsed:   25.6s remaining:    7.7s\n",
      "[Parallel(n_jobs=-1)]: Done  30 out of  30 | elapsed:   32.8s finished\n",
      "[Parallel(n_jobs=12)]: Using backend ThreadingBackend with 12 concurrent workers.\n",
      "[Parallel(n_jobs=12)]: Done  23 out of  30 | elapsed:    0.4s remaining:    0.1s\n",
      "[Parallel(n_jobs=12)]: Done  30 out of  30 | elapsed:    0.6s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RFC Accuracy: 0.9407237065919974\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x1be9c16cf90>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Load and preprocess your data (X_train, y_train, X_test, y_test)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.2, train_size=0.8)\n",
    "\n",
    "#We add a weight to the classes to prioritize precision in favor of class 1 \n",
    "class_weights = {0:1, 1:55}\n",
    "rfc_model = RFC(n_estimators=30,max_depth=40,min_samples_leaf=10,min_samples_split=10,n_jobs=-1,class_weight=class_weights,verbose=2, random_state=0)\n",
    "\n",
    "# Train the model\n",
    "rfc_model.fit(X_train, y_train)\n",
    "\n",
    "#We adjust the decision threshold to make the model more conservative in predicting class 0. By lowering the threshold, you require a much higher level of confidence for a sample to be classified as class 0, reducing the risk of false negatives.\n",
    "threshold = 0.01  # Adjust the threshold as needed\n",
    "y_pred_prob = rfc_model.predict_proba(X_test)[:, 1]\n",
    "y_pred = (y_pred_prob > threshold).astype(int)\n",
    "\n",
    "\n",
    "accuracy_rfc = accuracy_score(y_test, y_pred)\n",
    "print(\"RFC Accuracy:\", accuracy_rfc)\n",
    "\n",
    "cm_rfc = confusion_matrix(y_test, y_pred)\n",
    "disp_rfc = ConfusionMatrixDisplay(cm_rfc)\n",
    "disp_rfc.plot()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "1417f328e3a7d27b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-01-24T22:02:56.704654900Z",
     "start_time": "2024-01-24T22:02:56.619728700Z"
    },
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "disp_rfc.figure_.savefig(\"./Img/RFC_FinalResult0_18k.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "b5fc5f52e17b6ed2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-01-24T22:03:18.513292900Z",
     "start_time": "2024-01-24T22:03:17.712041600Z"
    },
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9407237065919974\n",
      "Precision: 0.9261106240698077\n",
      "Recall: 1.0\n"
     ]
    }
   ],
   "source": [
    "accuracy = accuracy_score(y_test, y_pred)\n",
    "precision = precision_score(y_test, y_pred)\n",
    "recall = recall_score(y_test, y_pred)\n",
    "\n",
    "print(\"Accuracy:\", accuracy)\n",
    "print(\"Precision:\", precision)\n",
    "print(\"Recall:\", recall)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e3836dbb8f5e60ca",
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}