speaker_change_detection_using_1_class_SVM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 383,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import sklearn  as sk\n",
    "from sklearn.model_selection import RepeatedKFold \n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import svm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# reading input wav file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 384,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "from python_speech_features import mfcc\n",
    "from python_speech_features import logfbank\n",
    "import scipy.io.wavfile as wav"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 385,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "numpy.ndarray"
      ]
     },
     "execution_count": 385,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(rate,sig) = wav.read(\"/home/nikita/Documents/Speaker change detection/data/LDC2007S10.wav\")\n",
    "data = mfcc(sig,rate)\n",
    "type(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 386,
   "metadata": {},
   "outputs": [],
   "source": [
    "#deleting energy coefficient from mfcc feature\n",
    "data = np.delete(data, 0, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 387,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "no of data samples= 179999\n",
      "(179999, 12)\n"
     ]
    }
   ],
   "source": [
    "no_of_data_samples=data.shape[0]\n",
    "print(\"no of data samples=\",no_of_data_samples)\n",
    "print(data.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reading actual change points found using transcript file(Dataset:NIST RT03 meeting data) [change_points_list.npy]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 388,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total no of change points= 448\n"
     ]
    }
   ],
   "source": [
    "\n",
    "cp=np.load(\"change_points_list.npy\")\n",
    "cp=cp.astype(int)\n",
    "# print(cp[0:100])\n",
    "no_of_change_points=cp.shape\n",
    "no_of_change_points=no_of_change_points[0]\n",
    "print(\"Total no of change points=\",no_of_change_points)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 389,
   "metadata": {},
   "outputs": [],
   "source": [
    "classifier=svm.OneClassSVM(nu=0.1,kernel=\"linear\",gamma=0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1classsvm.decision_function-Signed distance to the separating hyperplane.Signed distance is positive for an inlier and negative for an outlier."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 391,
   "metadata": {},
   "outputs": [],
   "source": [
    "speaker_accuracy=pd.DataFrame(columns=['data1','data2','data3','actual_speaker','next_speaker'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualizing 1classSVM performance by modeling 1classSVM on one speaker data and testing accuracy of next speaker on that model and that speaker data itself"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 392,
   "metadata": {},
   "outputs": [],
   "source": [
    "# speaker1 speaker2\n",
    "a=0\n",
    "t=cp[0]\n",
    "data_sample=data[a:t]\n",
    "no_of_data_points,no_of_features=data_sample.shape\n",
    "#fitting 1classsvm to 1st speaker\n",
    "classifier.fit(data_sample,(no_of_data_points,12))\n",
    "#calculating signed distance for two speakers with respect to 1classsvm of 1st speaker\n",
    "p1=classifier.decision_function(data[0:cp[0]])\n",
    "pred1=classifier.predict(data[0:cp[0]])\n",
    "# print(p1.shape)\n",
    "p2=classifier.decision_function(data[cp[0]:cp[1]])\n",
    "pred2=classifier.predict(data[cp[0]:cp[1]])\n",
    "# print(p1.max())\n",
    "# print(p1.min())\n",
    "i1=p1.shape\n",
    "k1=0\n",
    "c1=0\n",
    "for k1 in range(0,i1[0]):\n",
    "    if(p1[k1]>0):\n",
    "        c1=c1+1\n",
    "#     print(c1)\n",
    "acc1=c1/i1[0]\n",
    "\n",
    "i2=p2.shape\n",
    "# print(i2)\n",
    "k2=0\n",
    "c2=0\n",
    "for k2 in range(0,i2[0]):\n",
    "    if(p2[k2]>0):\n",
    "        c2=c2+1\n",
    "#     print(c2)\n",
    "acc2=c2/i2[0]\n",
    "speaker_accuracy=speaker_accuracy.append({'data1':0, 'data2':cp[0], 'data3':cp[1], 'actual_speaker':acc1, 'next_speaker':acc2}, ignore_index=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Figure depicting 1-class SVM on trained data of single speaker only\n",
    "+1.0 represents correct classification \n",
    "\n",
    "-1.0  represents misclassification (very few)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 394,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fe176b5a940>"
      ]
     },
     "execution_count": 394,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2880x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 0 2750 3108 <-speaker change occurs like that in data point\n",
    "# so we modeled 1classSVM on 0-2749 data points and tested on 0-2749(Actual speaker) and 2750-3107(next speaker)\n",
    "# plotting predicted value (-1:data point does not belong to model,1:data point belong to model)\n",
    "y=np.arange(3108)\n",
    "fig = plt.figure()\n",
    "plt.figure(figsize=(40,3))\n",
    "ax = plt.subplot(111)\n",
    "ax.scatter(y[0:2750], pred1, color='r')\n",
    "# ax.scatter(y[2750:3025], pred2[0:275], color='b')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#  Figure depicting the change between  two speakers \n",
    "Blue=First Speaker (label +1.0 on y axis)\n",
    "\n",
    "Orange =Second Speaker (label -1.0 on y axis)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 395,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe176b37048>]"
      ]
     },
     "execution_count": 395,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 2880x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y=np.arange(3108)\n",
    "fig = plt.figure()\n",
    "plt.figure(figsize=(40,3))\n",
    "ax = plt.subplot(111)\n",
    "# print(pred2.shape)\n",
    "ax.plot(y[0:2750], pred1, '-o') #blue curve shows testing on actual speaker data\n",
    "ax.plot(y[2750:3108], pred2[0:358], '-o') #orange curve shows testing on next speaker data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#In the above graph blue points on -1.0 label shows misclassification of speaker1 data and orange points\n",
    "# on +1.0 label shows misclassification of speaker2 data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 398,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "ind=0\n",
    "for ind in range(0,75): \n",
    "    a=cp[ind]\n",
    "    t=cp[ind+1]\n",
    "    data_sample=data[a:t]\n",
    "    no_of_data_points,no_of_features=data_sample.shape\n",
    "    #fitting 1classsvm to 1st speaker\n",
    "    classifier.fit(data_sample,(no_of_data_points,12))\n",
    "    #calculating signed distance for two speakers with respect to 1classsvm of 1st speaker\n",
    "    p1=classifier.decision_function(data[cp[ind]:cp[ind+1]])\n",
    "    p2=classifier.decision_function(data[cp[ind+1]:cp[ind+2]])\n",
    "    pred11=classifier.predict(data[cp[ind]:cp[ind+1]])\n",
    "    pred22=classifier.predict(data[cp[ind+1]:cp[ind+2]])\n",
    "    i1=p1.shape\n",
    "    k1=0\n",
    "    c1=0\n",
    "    for k1 in range(0,i1[0]):\n",
    "        if(p1[k1]>0):\n",
    "            c1=c1+1\n",
    "#     print(c1)\n",
    "    acc1=c1/i1[0]\n",
    "\n",
    "    i2=p2.shape\n",
    "    # print(i2)\n",
    "    k2=0\n",
    "    c2=0\n",
    "    for k2 in range(0,i2[0]):\n",
    "        if(p2[k2]>0):\n",
    "            c2=c2+1\n",
    "#     print(c2)\n",
    "    acc2=c2/i2[0]\n",
    "    speaker_accuracy=speaker_accuracy.append({'data1':cp[ind], 'data2':cp[ind+1], 'data3':cp[ind+2], 'actual_speaker':acc1, 'next_speaker':acc2}, ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 401,
   "metadata": {},
   "outputs": [
    {
     "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>data1</th>\n",
       "      <th>data2</th>\n",
       "      <th>data3</th>\n",
       "      <th>actual_speaker</th>\n",
       "      <th>next_speaker</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>2750.0</td>\n",
       "      <td>3108.0</td>\n",
       "      <td>0.899636</td>\n",
       "      <td>0.321229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2750.0</td>\n",
       "      <td>3108.0</td>\n",
       "      <td>3807.0</td>\n",
       "      <td>0.896648</td>\n",
       "      <td>0.836910</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3108.0</td>\n",
       "      <td>3807.0</td>\n",
       "      <td>8264.0</td>\n",
       "      <td>0.898426</td>\n",
       "      <td>0.743774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3807.0</td>\n",
       "      <td>8264.0</td>\n",
       "      <td>8382.0</td>\n",
       "      <td>0.899933</td>\n",
       "      <td>0.093220</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>8264.0</td>\n",
       "      <td>8382.0</td>\n",
       "      <td>13075.0</td>\n",
       "      <td>0.889831</td>\n",
       "      <td>0.548476</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>8382.0</td>\n",
       "      <td>13075.0</td>\n",
       "      <td>13742.0</td>\n",
       "      <td>0.900064</td>\n",
       "      <td>0.673163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>13075.0</td>\n",
       "      <td>13742.0</td>\n",
       "      <td>17372.0</td>\n",
       "      <td>0.904048</td>\n",
       "      <td>0.936639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>13742.0</td>\n",
       "      <td>17372.0</td>\n",
       "      <td>18906.0</td>\n",
       "      <td>0.900275</td>\n",
       "      <td>0.672099</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>17372.0</td>\n",
       "      <td>18906.0</td>\n",
       "      <td>21428.0</td>\n",
       "      <td>0.900913</td>\n",
       "      <td>0.688739</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>18906.0</td>\n",
       "      <td>21428.0</td>\n",
       "      <td>21473.0</td>\n",
       "      <td>0.899683</td>\n",
       "      <td>0.822222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>21428.0</td>\n",
       "      <td>21473.0</td>\n",
       "      <td>24408.0</td>\n",
       "      <td>0.911111</td>\n",
       "      <td>0.268484</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>21473.0</td>\n",
       "      <td>24408.0</td>\n",
       "      <td>26444.0</td>\n",
       "      <td>0.899830</td>\n",
       "      <td>0.730354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>24408.0</td>\n",
       "      <td>26444.0</td>\n",
       "      <td>27035.0</td>\n",
       "      <td>0.899804</td>\n",
       "      <td>0.903553</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>26444.0</td>\n",
       "      <td>27035.0</td>\n",
       "      <td>27138.0</td>\n",
       "      <td>0.901861</td>\n",
       "      <td>0.135922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>27035.0</td>\n",
       "      <td>27138.0</td>\n",
       "      <td>27764.0</td>\n",
       "      <td>0.902913</td>\n",
       "      <td>0.407348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>27138.0</td>\n",
       "      <td>27764.0</td>\n",
       "      <td>28607.0</td>\n",
       "      <td>0.899361</td>\n",
       "      <td>0.441281</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>27764.0</td>\n",
       "      <td>28607.0</td>\n",
       "      <td>30902.0</td>\n",
       "      <td>0.899170</td>\n",
       "      <td>0.212200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>28607.0</td>\n",
       "      <td>30902.0</td>\n",
       "      <td>31641.0</td>\n",
       "      <td>0.901525</td>\n",
       "      <td>0.522327</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>30902.0</td>\n",
       "      <td>31641.0</td>\n",
       "      <td>32136.0</td>\n",
       "      <td>0.899865</td>\n",
       "      <td>0.800000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>31641.0</td>\n",
       "      <td>32136.0</td>\n",
       "      <td>32878.0</td>\n",
       "      <td>0.905051</td>\n",
       "      <td>0.632075</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>32136.0</td>\n",
       "      <td>32878.0</td>\n",
       "      <td>35843.0</td>\n",
       "      <td>0.898922</td>\n",
       "      <td>0.632715</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>32878.0</td>\n",
       "      <td>35843.0</td>\n",
       "      <td>38778.0</td>\n",
       "      <td>0.900169</td>\n",
       "      <td>0.777853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>35843.0</td>\n",
       "      <td>38778.0</td>\n",
       "      <td>41504.0</td>\n",
       "      <td>0.899489</td>\n",
       "      <td>0.872340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>38778.0</td>\n",
       "      <td>41504.0</td>\n",
       "      <td>42352.0</td>\n",
       "      <td>0.899486</td>\n",
       "      <td>0.522406</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>41504.0</td>\n",
       "      <td>42352.0</td>\n",
       "      <td>45617.0</td>\n",
       "      <td>0.900943</td>\n",
       "      <td>0.339663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>42352.0</td>\n",
       "      <td>45617.0</td>\n",
       "      <td>46677.0</td>\n",
       "      <td>0.899541</td>\n",
       "      <td>0.585849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>45617.0</td>\n",
       "      <td>46677.0</td>\n",
       "      <td>49051.0</td>\n",
       "      <td>0.901887</td>\n",
       "      <td>0.699663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>46677.0</td>\n",
       "      <td>49051.0</td>\n",
       "      <td>52304.0</td>\n",
       "      <td>0.899747</td>\n",
       "      <td>0.829081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>49051.0</td>\n",
       "      <td>52304.0</td>\n",
       "      <td>52668.0</td>\n",
       "      <td>0.899785</td>\n",
       "      <td>0.892857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>52304.0</td>\n",
       "      <td>52668.0</td>\n",
       "      <td>53149.0</td>\n",
       "      <td>0.892857</td>\n",
       "      <td>0.918919</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>72774.0</td>\n",
       "      <td>85870.0</td>\n",
       "      <td>89323.0</td>\n",
       "      <td>0.899969</td>\n",
       "      <td>0.894584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>85870.0</td>\n",
       "      <td>89323.0</td>\n",
       "      <td>90057.0</td>\n",
       "      <td>0.900376</td>\n",
       "      <td>0.765668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>89323.0</td>\n",
       "      <td>90057.0</td>\n",
       "      <td>90878.0</td>\n",
       "      <td>0.900545</td>\n",
       "      <td>0.534714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>90057.0</td>\n",
       "      <td>90878.0</td>\n",
       "      <td>91927.0</td>\n",
       "      <td>0.901340</td>\n",
       "      <td>0.728313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>90878.0</td>\n",
       "      <td>91927.0</td>\n",
       "      <td>92861.0</td>\n",
       "      <td>0.900858</td>\n",
       "      <td>0.640257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>91927.0</td>\n",
       "      <td>92861.0</td>\n",
       "      <td>94709.0</td>\n",
       "      <td>0.899358</td>\n",
       "      <td>0.721861</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>92861.0</td>\n",
       "      <td>94709.0</td>\n",
       "      <td>95475.0</td>\n",
       "      <td>0.900433</td>\n",
       "      <td>0.933420</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>94709.0</td>\n",
       "      <td>95475.0</td>\n",
       "      <td>98252.0</td>\n",
       "      <td>0.900783</td>\n",
       "      <td>0.658264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>95475.0</td>\n",
       "      <td>98252.0</td>\n",
       "      <td>99045.0</td>\n",
       "      <td>0.899532</td>\n",
       "      <td>0.863808</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>98252.0</td>\n",
       "      <td>99045.0</td>\n",
       "      <td>99872.0</td>\n",
       "      <td>0.899117</td>\n",
       "      <td>0.896010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>99045.0</td>\n",
       "      <td>99872.0</td>\n",
       "      <td>100426.0</td>\n",
       "      <td>0.899637</td>\n",
       "      <td>0.770758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>99872.0</td>\n",
       "      <td>100426.0</td>\n",
       "      <td>103116.0</td>\n",
       "      <td>0.898917</td>\n",
       "      <td>0.743866</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>100426.0</td>\n",
       "      <td>103116.0</td>\n",
       "      <td>103236.0</td>\n",
       "      <td>0.898885</td>\n",
       "      <td>0.866667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>103116.0</td>\n",
       "      <td>103236.0</td>\n",
       "      <td>103489.0</td>\n",
       "      <td>0.900000</td>\n",
       "      <td>0.656126</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>103236.0</td>\n",
       "      <td>103489.0</td>\n",
       "      <td>105825.0</td>\n",
       "      <td>0.897233</td>\n",
       "      <td>0.584760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>103489.0</td>\n",
       "      <td>105825.0</td>\n",
       "      <td>105958.0</td>\n",
       "      <td>0.899829</td>\n",
       "      <td>0.947368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>105825.0</td>\n",
       "      <td>105958.0</td>\n",
       "      <td>106159.0</td>\n",
       "      <td>0.894737</td>\n",
       "      <td>0.850746</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>105958.0</td>\n",
       "      <td>106159.0</td>\n",
       "      <td>107732.0</td>\n",
       "      <td>0.895522</td>\n",
       "      <td>0.535919</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64</th>\n",
       "      <td>106159.0</td>\n",
       "      <td>107732.0</td>\n",
       "      <td>107814.0</td>\n",
       "      <td>0.899555</td>\n",
       "      <td>0.902439</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65</th>\n",
       "      <td>107732.0</td>\n",
       "      <td>107814.0</td>\n",
       "      <td>108992.0</td>\n",
       "      <td>0.902439</td>\n",
       "      <td>0.798812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66</th>\n",
       "      <td>107814.0</td>\n",
       "      <td>108992.0</td>\n",
       "      <td>109970.0</td>\n",
       "      <td>0.899830</td>\n",
       "      <td>0.574642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>108992.0</td>\n",
       "      <td>109970.0</td>\n",
       "      <td>110619.0</td>\n",
       "      <td>0.899796</td>\n",
       "      <td>0.667180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>109970.0</td>\n",
       "      <td>110619.0</td>\n",
       "      <td>125103.0</td>\n",
       "      <td>0.898305</td>\n",
       "      <td>0.702430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>110619.0</td>\n",
       "      <td>125103.0</td>\n",
       "      <td>129271.0</td>\n",
       "      <td>0.899820</td>\n",
       "      <td>0.907150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>125103.0</td>\n",
       "      <td>129271.0</td>\n",
       "      <td>129499.0</td>\n",
       "      <td>0.899712</td>\n",
       "      <td>0.894737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>129271.0</td>\n",
       "      <td>129499.0</td>\n",
       "      <td>131079.0</td>\n",
       "      <td>0.903509</td>\n",
       "      <td>0.572152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>129499.0</td>\n",
       "      <td>131079.0</td>\n",
       "      <td>131726.0</td>\n",
       "      <td>0.900000</td>\n",
       "      <td>0.823802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73</th>\n",
       "      <td>131079.0</td>\n",
       "      <td>131726.0</td>\n",
       "      <td>132817.0</td>\n",
       "      <td>0.901082</td>\n",
       "      <td>0.535289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74</th>\n",
       "      <td>131726.0</td>\n",
       "      <td>132817.0</td>\n",
       "      <td>135975.0</td>\n",
       "      <td>0.901925</td>\n",
       "      <td>0.877137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>132817.0</td>\n",
       "      <td>135975.0</td>\n",
       "      <td>136109.0</td>\n",
       "      <td>0.899620</td>\n",
       "      <td>0.985075</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>76 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       data1     data2     data3  actual_speaker  next_speaker\n",
       "0        0.0    2750.0    3108.0        0.899636      0.321229\n",
       "1     2750.0    3108.0    3807.0        0.896648      0.836910\n",
       "2     3108.0    3807.0    8264.0        0.898426      0.743774\n",
       "3     3807.0    8264.0    8382.0        0.899933      0.093220\n",
       "4     8264.0    8382.0   13075.0        0.889831      0.548476\n",
       "5     8382.0   13075.0   13742.0        0.900064      0.673163\n",
       "6    13075.0   13742.0   17372.0        0.904048      0.936639\n",
       "7    13742.0   17372.0   18906.0        0.900275      0.672099\n",
       "8    17372.0   18906.0   21428.0        0.900913      0.688739\n",
       "9    18906.0   21428.0   21473.0        0.899683      0.822222\n",
       "10   21428.0   21473.0   24408.0        0.911111      0.268484\n",
       "11   21473.0   24408.0   26444.0        0.899830      0.730354\n",
       "12   24408.0   26444.0   27035.0        0.899804      0.903553\n",
       "13   26444.0   27035.0   27138.0        0.901861      0.135922\n",
       "14   27035.0   27138.0   27764.0        0.902913      0.407348\n",
       "15   27138.0   27764.0   28607.0        0.899361      0.441281\n",
       "16   27764.0   28607.0   30902.0        0.899170      0.212200\n",
       "17   28607.0   30902.0   31641.0        0.901525      0.522327\n",
       "18   30902.0   31641.0   32136.0        0.899865      0.800000\n",
       "19   31641.0   32136.0   32878.0        0.905051      0.632075\n",
       "20   32136.0   32878.0   35843.0        0.898922      0.632715\n",
       "21   32878.0   35843.0   38778.0        0.900169      0.777853\n",
       "22   35843.0   38778.0   41504.0        0.899489      0.872340\n",
       "23   38778.0   41504.0   42352.0        0.899486      0.522406\n",
       "24   41504.0   42352.0   45617.0        0.900943      0.339663\n",
       "25   42352.0   45617.0   46677.0        0.899541      0.585849\n",
       "26   45617.0   46677.0   49051.0        0.901887      0.699663\n",
       "27   46677.0   49051.0   52304.0        0.899747      0.829081\n",
       "28   49051.0   52304.0   52668.0        0.899785      0.892857\n",
       "29   52304.0   52668.0   53149.0        0.892857      0.918919\n",
       "..       ...       ...       ...             ...           ...\n",
       "46   72774.0   85870.0   89323.0        0.899969      0.894584\n",
       "47   85870.0   89323.0   90057.0        0.900376      0.765668\n",
       "48   89323.0   90057.0   90878.0        0.900545      0.534714\n",
       "49   90057.0   90878.0   91927.0        0.901340      0.728313\n",
       "50   90878.0   91927.0   92861.0        0.900858      0.640257\n",
       "51   91927.0   92861.0   94709.0        0.899358      0.721861\n",
       "52   92861.0   94709.0   95475.0        0.900433      0.933420\n",
       "53   94709.0   95475.0   98252.0        0.900783      0.658264\n",
       "54   95475.0   98252.0   99045.0        0.899532      0.863808\n",
       "55   98252.0   99045.0   99872.0        0.899117      0.896010\n",
       "56   99045.0   99872.0  100426.0        0.899637      0.770758\n",
       "57   99872.0  100426.0  103116.0        0.898917      0.743866\n",
       "58  100426.0  103116.0  103236.0        0.898885      0.866667\n",
       "59  103116.0  103236.0  103489.0        0.900000      0.656126\n",
       "60  103236.0  103489.0  105825.0        0.897233      0.584760\n",
       "61  103489.0  105825.0  105958.0        0.899829      0.947368\n",
       "62  105825.0  105958.0  106159.0        0.894737      0.850746\n",
       "63  105958.0  106159.0  107732.0        0.895522      0.535919\n",
       "64  106159.0  107732.0  107814.0        0.899555      0.902439\n",
       "65  107732.0  107814.0  108992.0        0.902439      0.798812\n",
       "66  107814.0  108992.0  109970.0        0.899830      0.574642\n",
       "67  108992.0  109970.0  110619.0        0.899796      0.667180\n",
       "68  109970.0  110619.0  125103.0        0.898305      0.702430\n",
       "69  110619.0  125103.0  129271.0        0.899820      0.907150\n",
       "70  125103.0  129271.0  129499.0        0.899712      0.894737\n",
       "71  129271.0  129499.0  131079.0        0.903509      0.572152\n",
       "72  129499.0  131079.0  131726.0        0.900000      0.823802\n",
       "73  131079.0  131726.0  132817.0        0.901082      0.535289\n",
       "74  131726.0  132817.0  135975.0        0.901925      0.877137\n",
       "75  132817.0  135975.0  136109.0        0.899620      0.985075\n",
       "\n",
       "[76 rows x 5 columns]"
      ]
     },
     "execution_count": 401,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "speaker_accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Trying incremental approach of detecting change point using 1class SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 402,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "may be change point at 2754\n",
      "may be change point at 3154\n",
      "may be change point at 3481\n",
      "may be change point at 3788\n",
      "may be change point at 4110\n",
      "may be change point at 4418\n",
      "may be change point at 4747\n",
      "may be change point at 5067\n",
      "may be change point at 5390\n",
      "may be change point at 5694\n",
      "may be change point at 6031\n",
      "may be change point at 6357\n",
      "may be change point at 6659\n",
      "may be change point at 6978\n",
      "may be change point at 7279\n",
      "may be change point at 7594\n",
      "may be change point at 7947\n",
      "may be change point at 8259\n",
      "may be change point at 8579\n",
      "may be change point at 8890\n",
      "may be change point at 9196\n",
      "may be change point at 9501\n",
      "may be change point at 9801\n",
      "may be change point at 10121\n",
      "may be change point at 10438\n",
      "may be change point at 10764\n",
      "may be change point at 11073\n",
      "may be change point at 11383\n",
      "may be change point at 11710\n",
      "may be change point at 12024\n",
      "may be change point at 12324\n",
      "may be change point at 12648\n",
      "may be change point at 12960\n",
      "may be change point at 13271\n",
      "may be change point at 13603\n",
      "may be change point at 13915\n",
      "may be change point at 14215\n",
      "may be change point at 14523\n",
      "may be change point at 14837\n",
      "may be change point at 15144\n",
      "may be change point at 15461\n",
      "may be change point at 15775\n",
      "may be change point at 16156\n",
      "may be change point at 16461\n",
      "may be change point at 16779\n",
      "may be change point at 17080\n",
      "may be change point at 17380\n",
      "may be change point at 17689\n",
      "may be change point at 17990\n",
      "may be change point at 18308\n",
      "may be change point at 18612\n",
      "may be change point at 18912\n",
      "may be change point at 19212\n",
      "may be change point at 19518\n",
      "may be change point at 19829\n",
      "may be change point at 20129\n",
      "may be change point at 20431\n",
      "may be change point at 20747\n",
      "may be change point at 21058\n",
      "may be change point at 21371\n",
      "may be change point at 21671\n",
      "may be change point at 21984\n",
      "may be change point at 22285\n",
      "may be change point at 22651\n",
      "may be change point at 22966\n",
      "may be change point at 23294\n",
      "may be change point at 23605\n",
      "may be change point at 23940\n",
      "may be change point at 24258\n",
      "may be change point at 24562\n",
      "may be change point at 24864\n",
      "may be change point at 25173\n",
      "may be change point at 25561\n",
      "may be change point at 25863\n",
      "may be change point at 26170\n",
      "may be change point at 26475\n",
      "may be change point at 26790\n",
      "may be change point at 27104\n",
      "may be change point at 27415\n",
      "may be change point at 27715\n",
      "may be change point at 28042\n",
      "may be change point at 28377\n",
      "may be change point at 28678\n",
      "may be change point at 28990\n",
      "may be change point at 29290\n",
      "may be change point at 29599\n",
      "may be change point at 29905\n",
      "may be change point at 30230\n",
      "may be change point at 30561\n",
      "may be change point at 30863\n",
      "may be change point at 31165\n",
      "may be change point at 31466\n",
      "may be change point at 31770\n",
      "may be change point at 32072\n",
      "may be change point at 32388\n",
      "may be change point at 32696\n",
      "may be change point at 32996\n",
      "may be change point at 33305\n",
      "may be change point at 33619\n",
      "may be change point at 33966\n",
      "may be change point at 34276\n",
      "may be change point at 34587\n",
      "may be change point at 34904\n",
      "may be change point at 35210\n",
      "may be change point at 35528\n",
      "may be change point at 35836\n",
      "may be change point at 36324\n",
      "may be change point at 36630\n",
      "may be change point at 36940\n",
      "may be change point at 37244\n",
      "may be change point at 37561\n",
      "may be change point at 37873\n",
      "may be change point at 38193\n",
      "may be change point at 38501\n",
      "may be change point at 38802\n",
      "may be change point at 39107\n",
      "may be change point at 39410\n",
      "may be change point at 39711\n",
      "may be change point at 40027\n",
      "may be change point at 40341\n",
      "may be change point at 40685\n",
      "may be change point at 41022\n",
      "may be change point at 41339\n",
      "may be change point at 41642\n",
      "may be change point at 41942\n",
      "may be change point at 42255\n",
      "may be change point at 42611\n",
      "may be change point at 42933\n",
      "may be change point at 43233\n",
      "may be change point at 43536\n",
      "may be change point at 43836\n",
      "may be change point at 44136\n",
      "may be change point at 44462\n",
      "may be change point at 44764\n",
      "may be change point at 45093\n",
      "may be change point at 45409\n",
      "may be change point at 45738\n",
      "may be change point at 46085\n",
      "may be change point at 46385\n",
      "may be change point at 46704\n",
      "may be change point at 47005\n",
      "may be change point at 47311\n",
      "may be change point at 47611\n",
      "may be change point at 47912\n",
      "may be change point at 48214\n",
      "may be change point at 48555\n",
      "may be change point at 48869\n",
      "may be change point at 49197\n",
      "may be change point at 49658\n",
      "may be change point at 49965\n",
      "may be change point at 50312\n",
      "may be change point at 50645\n",
      "may be change point at 50956\n",
      "may be change point at 51269\n",
      "may be change point at 51574\n",
      "may be change point at 51885\n",
      "may be change point at 52185\n",
      "may be change point at 52542\n",
      "may be change point at 52870\n",
      "may be change point at 53234\n",
      "may be change point at 53720\n",
      "may be change point at 54039\n",
      "may be change point at 54365\n",
      "may be change point at 54671\n",
      "may be change point at 55104\n",
      "may be change point at 55408\n",
      "may be change point at 55714\n",
      "may be change point at 56028\n",
      "may be change point at 56328\n",
      "may be change point at 56634\n",
      "may be change point at 56947\n",
      "may be change point at 57248\n",
      "may be change point at 57552\n",
      "may be change point at 57868\n",
      "may be change point at 58186\n",
      "may be change point at 58493\n",
      "may be change point at 58805\n",
      "may be change point at 59105\n",
      "may be change point at 59408\n",
      "may be change point at 59711\n",
      "may be change point at 60041\n",
      "may be change point at 60357\n",
      "may be change point at 60674\n",
      "may be change point at 60975\n",
      "may be change point at 61387\n",
      "may be change point at 61737\n",
      "may be change point at 62088\n",
      "may be change point at 62397\n",
      "may be change point at 62701\n",
      "may be change point at 63019\n",
      "may be change point at 63330\n",
      "may be change point at 63630\n",
      "may be change point at 63953\n",
      "may be change point at 64253\n",
      "may be change point at 64557\n",
      "may be change point at 64857\n",
      "may be change point at 65190\n",
      "may be change point at 65544\n",
      "may be change point at 65882\n",
      "may be change point at 66193\n",
      "may be change point at 66493\n",
      "may be change point at 66798\n",
      "may be change point at 67106\n",
      "may be change point at 67415\n",
      "may be change point at 67752\n",
      "may be change point at 68130\n",
      "may be change point at 68436\n",
      "may be change point at 68736\n",
      "may be change point at 69100\n",
      "may be change point at 69425\n",
      "may be change point at 69751\n",
      "may be change point at 70051\n",
      "may be change point at 70355\n",
      "may be change point at 70655\n",
      "may be change point at 70955\n",
      "may be change point at 71284\n",
      "may be change point at 71591\n",
      "may be change point at 71892\n",
      "may be change point at 72243\n",
      "may be change point at 72566\n",
      "may be change point at 72873\n",
      "may be change point at 73182\n",
      "may be change point at 73482\n",
      "may be change point at 73798\n",
      "may be change point at 74107\n",
      "may be change point at 74415\n",
      "may be change point at 74730\n",
      "may be change point at 75035\n",
      "may be change point at 75338\n",
      "may be change point at 75651\n",
      "may be change point at 75957\n",
      "may be change point at 76257\n",
      "may be change point at 76560\n",
      "may be change point at 76873\n",
      "may be change point at 77174\n",
      "may be change point at 77486\n",
      "may be change point at 77794\n",
      "may be change point at 78099\n",
      "may be change point at 78452\n",
      "may be change point at 78758\n",
      "may be change point at 79065\n",
      "may be change point at 79365\n",
      "may be change point at 79667\n",
      "may be change point at 79971\n",
      "may be change point at 80333\n",
      "may be change point at 80634\n",
      "may be change point at 80939\n",
      "may be change point at 81466\n",
      "may be change point at 81766\n",
      "may be change point at 82081\n",
      "may be change point at 82381\n",
      "may be change point at 82681\n",
      "may be change point at 83045\n",
      "may be change point at 83371\n",
      "may be change point at 83742\n",
      "may be change point at 84047\n",
      "may be change point at 84380\n",
      "may be change point at 84680\n",
      "may be change point at 84992\n",
      "may be change point at 85331\n",
      "may be change point at 85761\n",
      "may be change point at 86062\n",
      "may be change point at 86362\n",
      "may be change point at 86673\n",
      "may be change point at 87002\n",
      "may be change point at 87302\n",
      "may be change point at 87615\n",
      "may be change point at 87986\n",
      "may be change point at 88335\n",
      "may be change point at 88638\n",
      "may be change point at 88961\n",
      "may be change point at 89264\n",
      "may be change point at 89584\n",
      "may be change point at 89914\n",
      "may be change point at 90244\n",
      "may be change point at 90595\n",
      "may be change point at 90896\n",
      "may be change point at 91251\n",
      "may be change point at 91621\n",
      "may be change point at 91921\n",
      "may be change point at 92236\n",
      "may be change point at 92538\n",
      "may be change point at 92839\n",
      "may be change point at 93151\n",
      "may be change point at 93452\n",
      "may be change point at 93752\n",
      "may be change point at 94065\n",
      "may be change point at 94369\n",
      "may be change point at 94683\n",
      "may be change point at 94998\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "may be change point at 95370\n",
      "may be change point at 95695\n",
      "may be change point at 96001\n",
      "may be change point at 96303\n",
      "may be change point at 96605\n",
      "may be change point at 96913\n",
      "may be change point at 97231\n",
      "may be change point at 97590\n",
      "may be change point at 97959\n",
      "may be change point at 98307\n",
      "may be change point at 98632\n",
      "may be change point at 98942\n",
      "may be change point at 99271\n",
      "may be change point at 99580\n",
      "may be change point at 99883\n",
      "may be change point at 100217\n",
      "may be change point at 100517\n",
      "may be change point at 100818\n",
      "may be change point at 101118\n",
      "may be change point at 101422\n",
      "may be change point at 101732\n",
      "may be change point at 102041\n",
      "may be change point at 102357\n",
      "may be change point at 102686\n",
      "may be change point at 102995\n",
      "may be change point at 103295\n",
      "may be change point at 103617\n",
      "may be change point at 103933\n",
      "may be change point at 104236\n",
      "may be change point at 104552\n",
      "may be change point at 104874\n",
      "may be change point at 105184\n",
      "may be change point at 105500\n",
      "may be change point at 105811\n",
      "may be change point at 106144\n",
      "may be change point at 106446\n",
      "may be change point at 106754\n",
      "may be change point at 107081\n",
      "may be change point at 107429\n",
      "may be change point at 107732\n",
      "may be change point at 108058\n",
      "may be change point at 108358\n",
      "may be change point at 108668\n",
      "may be change point at 108973\n",
      "may be change point at 109285\n",
      "may be change point at 109588\n",
      "may be change point at 109891\n",
      "may be change point at 110191\n",
      "may be change point at 110496\n",
      "may be change point at 110806\n",
      "may be change point at 111106\n",
      "may be change point at 111406\n",
      "may be change point at 111746\n",
      "may be change point at 112051\n",
      "may be change point at 112356\n",
      "may be change point at 112656\n",
      "may be change point at 112998\n",
      "may be change point at 113340\n",
      "may be change point at 113641\n",
      "may be change point at 113965\n",
      "may be change point at 114279\n",
      "may be change point at 114694\n",
      "may be change point at 115013\n",
      "may be change point at 115320\n",
      "may be change point at 115622\n",
      "may be change point at 115922\n",
      "may be change point at 116231\n",
      "may be change point at 116531\n",
      "may be change point at 116854\n",
      "may be change point at 117156\n",
      "may be change point at 117458\n",
      "may be change point at 117758\n",
      "may be change point at 118116\n",
      "may be change point at 118431\n",
      "may be change point at 118752\n",
      "may be change point at 119060\n",
      "may be change point at 119381\n",
      "may be change point at 119691\n",
      "may be change point at 119991\n",
      "may be change point at 120300\n",
      "may be change point at 120606\n",
      "may be change point at 120906\n",
      "may be change point at 121212\n",
      "may be change point at 121585\n",
      "may be change point at 121886\n",
      "may be change point at 122227\n",
      "may be change point at 122552\n",
      "may be change point at 122870\n",
      "may be change point at 123174\n",
      "may be change point at 123483\n",
      "may be change point at 123783\n",
      "may be change point at 124083\n",
      "may be change point at 124400\n",
      "may be change point at 124771\n",
      "may be change point at 125073\n",
      "may be change point at 125384\n",
      "may be change point at 125688\n",
      "may be change point at 125992\n",
      "may be change point at 126304\n",
      "may be change point at 126622\n",
      "may be change point at 126922\n",
      "may be change point at 127227\n",
      "may be change point at 127527\n",
      "may be change point at 127828\n",
      "may be change point at 128128\n",
      "may be change point at 128430\n",
      "may be change point at 128753\n",
      "may be change point at 129067\n",
      "may be change point at 129438\n",
      "may be change point at 129748\n",
      "may be change point at 130049\n",
      "may be change point at 130352\n",
      "may be change point at 130654\n",
      "may be change point at 130961\n",
      "may be change point at 131274\n",
      "may be change point at 131652\n",
      "may be change point at 131986\n",
      "may be change point at 132290\n",
      "may be change point at 132617\n",
      "may be change point at 132922\n",
      "may be change point at 133222\n",
      "may be change point at 133522\n",
      "may be change point at 133858\n",
      "may be change point at 134175\n",
      "may be change point at 134485\n",
      "may be change point at 134792\n",
      "may be change point at 135092\n",
      "may be change point at 135393\n",
      "may be change point at 135736\n",
      "may be change point at 136044\n",
      "may be change point at 136345\n",
      "may be change point at 136646\n",
      "may be change point at 136971\n",
      "may be change point at 137275\n",
      "may be change point at 137612\n",
      "may be change point at 137932\n",
      "may be change point at 138238\n",
      "may be change point at 138551\n",
      "may be change point at 138851\n",
      "may be change point at 139176\n",
      "may be change point at 139485\n",
      "may be change point at 139792\n",
      "may be change point at 140102\n",
      "may be change point at 140408\n",
      "may be change point at 140728\n",
      "may be change point at 141032\n",
      "may be change point at 141342\n",
      "may be change point at 141642\n",
      "may be change point at 141948\n",
      "may be change point at 142261\n",
      "may be change point at 142590\n",
      "may be change point at 142892\n",
      "may be change point at 143243\n",
      "may be change point at 143570\n",
      "may be change point at 143893\n",
      "may be change point at 144197\n",
      "may be change point at 144499\n",
      "may be change point at 144808\n",
      "may be change point at 145114\n",
      "may be change point at 145416\n",
      "may be change point at 145762\n",
      "may be change point at 146064\n",
      "may be change point at 146384\n",
      "may be change point at 146731\n",
      "may be change point at 147119\n",
      "may be change point at 147489\n",
      "may be change point at 147789\n",
      "may be change point at 148125\n",
      "may be change point at 148495\n",
      "may be change point at 148797\n",
      "may be change point at 149098\n",
      "may be change point at 149403\n",
      "may be change point at 149705\n",
      "may be change point at 150005\n",
      "may be change point at 150305\n",
      "may be change point at 150607\n",
      "may be change point at 150932\n",
      "may be change point at 151364\n",
      "may be change point at 151664\n",
      "may be change point at 151965\n",
      "may be change point at 152265\n",
      "may be change point at 152675\n",
      "may be change point at 152985\n",
      "may be change point at 153286\n",
      "may be change point at 153609\n",
      "may be change point at 153919\n",
      "may be change point at 154241\n",
      "may be change point at 154541\n",
      "may be change point at 154842\n",
      "may be change point at 155152\n",
      "may be change point at 155459\n",
      "may be change point at 155830\n",
      "may be change point at 156133\n",
      "may be change point at 156436\n",
      "may be change point at 156750\n",
      "may be change point at 157084\n",
      "may be change point at 157384\n",
      "may be change point at 157699\n",
      "may be change point at 157999\n",
      "may be change point at 158314\n",
      "may be change point at 158635\n",
      "may be change point at 158954\n",
      "may be change point at 159279\n",
      "may be change point at 159586\n",
      "may be change point at 159926\n",
      "may be change point at 160226\n",
      "may be change point at 160538\n",
      "may be change point at 160845\n",
      "may be change point at 161146\n",
      "may be change point at 161456\n",
      "may be change point at 161774\n",
      "may be change point at 162074\n",
      "may be change point at 162386\n",
      "may be change point at 162686\n",
      "may be change point at 163003\n",
      "may be change point at 163378\n",
      "may be change point at 163685\n",
      "may be change point at 163992\n",
      "may be change point at 164321\n",
      "may be change point at 164630\n",
      "may be change point at 164982\n",
      "may be change point at 165282\n",
      "may be change point at 165596\n",
      "may be change point at 165906\n",
      "may be change point at 166334\n",
      "may be change point at 166691\n",
      "may be change point at 166994\n",
      "may be change point at 167294\n",
      "may be change point at 167603\n",
      "may be change point at 167903\n",
      "may be change point at 168230\n",
      "may be change point at 168531\n",
      "may be change point at 168851\n",
      "may be change point at 169181\n",
      "may be change point at 169513\n",
      "may be change point at 169813\n",
      "may be change point at 170129\n",
      "may be change point at 170429\n",
      "may be change point at 170756\n",
      "may be change point at 171057\n",
      "may be change point at 171410\n",
      "may be change point at 171754\n",
      "may be change point at 172063\n",
      "may be change point at 172365\n",
      "may be change point at 172705\n",
      "may be change point at 173044\n",
      "may be change point at 173372\n",
      "may be change point at 173673\n",
      "may be change point at 174003\n",
      "may be change point at 174308\n",
      "may be change point at 174610\n",
      "may be change point at 174913\n",
      "may be change point at 175216\n",
      "may be change point at 175535\n",
      "may be change point at 175850\n",
      "may be change point at 176342\n",
      "may be change point at 176986\n",
      "may be change point at 177286\n",
      "may be change point at 177603\n",
      "may be change point at 177913\n",
      "may be change point at 178216\n",
      "may be change point at 178527\n",
      "may be change point at 178837\n",
      "may be change point at 179144\n",
      "may be change point at 179452\n",
      "may be change point at 179758\n"
     ]
    }
   ],
   "source": [
    "#c = no of change point detected\n",
    "c=0\n",
    "i=2750\n",
    "classifier.fit(data[0:2750],(2750,12))\n",
    "while i<no_of_data_samples:\n",
    "    if(classifier.predict([data[i]])==-1):\n",
    "        #data point does not belong to cluster, make new svm\n",
    "        print(\"may be change point at\",i)\n",
    "        a=i\n",
    "        t=a+300\n",
    "        if(i+300<no_of_data_samples):\n",
    "            data_sample=data[a:t]\n",
    "            classifier.fit(data_sample,(300,12))\n",
    "        i=i+300\n",
    "        c=c+1\n",
    "    else:\n",
    "        i=i+1\n",
    "        \n",
    "  #  classifier.fit(data_sample,(t-a,13))\n",
    "       \n",
    "    #else:\n",
    "        #print(\"1\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 403,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "no of change points predicted= 556\n"
     ]
    }
   ],
   "source": [
    "print(\"no of change points predicted=\",c)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Conclusion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# This notebook deals with detecting speaker change points using 1 class svm. "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}