4-Real-time_prediction_with_concept_drift_adaptation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# A Lightweight Concept Drift Detection and Adaptation Framework for IoT Data Streams \n",
    "This is the code for the paper entitled \"**A Lightweight Concept Drift Detection and Adaptation Framework for IoT Data Streams**\" accepted in IEEE Internet of Things Magazine.  \n",
    "Authors: Li Yang (lyang339@uwo.ca) and Abdallah Shami (Abdallah.Shami@uwo.ca)  \n",
    "Organization: The Optimized Computing and Communications (OC2) Lab, ECE Department, Western University\n",
    "\n",
    "**Notebook 4: Real-time prediction with concept drift adaptation**  \n",
    "Aims:  \n",
    "&nbsp; 1): Develop a real-time classifier with concept drift detection  \n",
    "&nbsp; 2): Plot a figure/gif to show the real-time prediction results that can be updated over time."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "import lightgbm as lgb\n",
    "from sklearn.metrics import accuracy_score\n",
    "from river import metrics\n",
    "from river import stream\n",
    "import matplotlib.pyplot as plt\n",
    "import statistics \n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read the NSL-KDD dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"NSL_KDD_binary(train+test).csv\")\n",
    "df1 = pd.read_csv(\"NSL_KDD_binary_train.csv\")\n",
    "df2 = pd.read_csv(\"NSL_KDD_binary_test.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data sampling\n",
    "Use the last 10% of training set and the entire test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df_used = int(len(df1)*0.1) + len(df2)\n",
    "df0 = df.iloc[-df_used:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train-test split\n",
    "10% training set, and 90% test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = df0.drop(['label'],axis=1)\n",
    "y = df0['label']\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,y, train_size = 0.1, test_size = 0.9, shuffle=False, random_state = 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Offline learning: LightGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8420969424858507\n"
     ]
    }
   ],
   "source": [
    "classifier = lgb.LGBMClassifier()\n",
    "# classifier = joblib.load('Optimized_lightGBM.pkl') \n",
    "classifier.fit(X_train,y_train)\n",
    "predictions = classifier.predict(X_test)\n",
    "print(\"Accuracy: \"+str(accuracy_score(y_test,predictions)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib qt5\n",
    "#%matplotlib inline\n",
    "\n",
    "def oasw_real_time(a=0.99, b=0.95, win1=200, win2=1000):\n",
    "    a = round(a, 3)\n",
    "    b = round(b, 3)\n",
    "    win1 = round(win1)\n",
    "    win2 = round(win2)\n",
    "\n",
    "    plt.ion() #start interactive mode \n",
    "    #plt.figure(1, constrained_layout=True)\n",
    "    plt.figure(1,figsize=(12,7.5))\n",
    "    sns.set_style(\"darkgrid\")\n",
    "    \n",
    "    metric = metrics.Accuracy()\n",
    "    metric2 = metrics.Accuracy()\n",
    "\n",
    "    i = 0\n",
    "    t = []\n",
    "    yt = []\n",
    "    yp = []\n",
    "    m = []\n",
    "    m2 = []\n",
    "    x_new=[]\n",
    "    y_new=[]\n",
    "    dr=[0]\n",
    "    d=0\n",
    "    f=0\n",
    "    tt=0\n",
    "    th=0\n",
    "    xt=[]\n",
    "\n",
    "    classifier = lgb.LGBMClassifier() # Train the initial LightGBM model\n",
    "    classifier.fit(X_train,y_train)\n",
    "    \n",
    "    classifier2 = lgb.LGBMClassifier() # Train an offline LightGBM model as a comparison model\n",
    "    classifier2.fit(X_train,y_train)\n",
    "\n",
    "    for xi, yi in stream.iter_pandas(X_test, y_test):\n",
    "\n",
    "        xi2=np.array(list(xi.values()))\n",
    "        y_pred = classifier.predict(xi2.reshape(1, -1))      # make a prediction\n",
    "        metric = metric.update(yi, y_pred)  # update the metric\n",
    "\n",
    "        y_pred2= classifier2.predict(xi2.reshape(1, -1)) \n",
    "        metric2 = metric2.update(yi, y_pred2)\n",
    "        \n",
    "        # Store the y_test, y_pred, x_test, and real-time accuracy\n",
    "        t.append(i)\n",
    "        m.append(metric.get()*100)\n",
    "        yt.append(yi)\n",
    "        yp.append(y_pred)\n",
    "        m2.append(metric2.get()*100)\n",
    "        xt.append(xi2)\n",
    "\n",
    "        # Monitor the accuracy changes in the sliding window\n",
    "        if i>2*win1:\n",
    "            acc1=accuracy_score(yt[i-win1:],yp[i-win1:]) # Current window accuracy\n",
    "            acc2=accuracy_score(yt[i-2*win1:i-win1],yp[i-2*win1:i-win1]) # Last window accuracy\n",
    "            if (d==0)&(acc1<a*acc2): # If the window accuracy drops to the warning level\n",
    "                x_new.append(xi2)\n",
    "                y_new.append(yi)\n",
    "                d=1\n",
    "            if d==1: # In the warning level\n",
    "                tt=len(y_new)\n",
    "                if acc1<b*acc2: # If the window accuracy drops to the drift level\n",
    "                    dr.append(i)# Record the drift start point\n",
    "                    f=i\n",
    "                    if tt<win1: # if enough new concept samples are collected\n",
    "                        classifier.fit(xt[i-win1:],yt[i-win1:]) \n",
    "                    else:\n",
    "                        classifier.fit(x_new,y_new)\n",
    "                    d=2\n",
    "                elif (acc1>a*acc2)|(tt==win2): # If the window accuracy increases back to the normal level (false alarm)\n",
    "                    x_new=[]\n",
    "                    y_new=[]\n",
    "                    d=0\n",
    "                else:\n",
    "                    x_new.append(xi2)\n",
    "                    y_new.append(yi)\n",
    "            if d==2: # In the drift level\n",
    "                tt=len(y_new)\n",
    "                acc3=accuracy_score(yt[f:f+win1],yp[f:f+win1])\n",
    "                x_new.append(xi2)\n",
    "                y_new.append(yi)\n",
    "                if tt>=win1:\n",
    "                    if (acc1<a*acc3): # When new concept accuracy drops to the warning level\n",
    "                        if th==0:\n",
    "                            classifier.fit(x_new,y_new) # Retrain the classifier on all the newly collected samples to obtain a robust classifier\n",
    "                            th=1\n",
    "                    if (th==1)&(tt==win2): # When sufficient new concept samples are collected\n",
    "                        classifier.fit(x_new,y_new) #obtain a robust classifier\n",
    "                        x_new=[]\n",
    "                        y_new=[]\n",
    "                        d=0 # Go back to the normal state for next potential drift detection\n",
    "                        th=0\n",
    "\n",
    "        \n",
    "#     if i == (10000):\n",
    "        plt.clf() \n",
    "        plt.plot(t,m,'-b',label='RMSE with drift')\n",
    "        plt.plot(t,m2,'pink',label='RMSE without drift')\n",
    "        #plt.plot(t,yp,'-g',label='True values')\n",
    "        #plt.plot(t,yt,'-r',label='Predicted values')\n",
    "        plt.legend(loc='upper left')\n",
    "        plt.title('NSL-KDD Dataset Online Learning Results', fontsize=15,pad =10)\n",
    "        plt.suptitle('NSL-KDD Dataset Results'+'\\nAvg Test Accuracy of Adaptive LightGBM: '+str(round(metric.get()*100, 2))+'\\nAvg Test Accuracy of Offline LightGBM: '\n",
    "                     +str(round(metric2.get()*100, 2))+'\\nDrift detected at timestamps: '+str(dr[1:]), fontsize=10,verticalalignment='baseline')\n",
    "        plt.xlabel('Number of samples')\n",
    "        plt.ylabel('Accuracy (%)')\n",
    "\n",
    "        plt.draw()\n",
    "\n",
    "        plt.pause(0.000000001)\n",
    "\n",
    "        i=i+1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x540 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "oasw_real_time(a = 0.978, b = 0.931, win1 = 611, win2 = 4320)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "tf36",
   "language": "python",
   "name": "tf36"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}