classification.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:246ea530dfb22efc00dbd8e88e4576d34ee34957cb1dcf5bd5bdc495ca71b027"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numpy as np\n",
      "import scipy as sp\n",
      "import pandas as pd\n",
      "import matplotlib.pyplot as plt\n",
      "#import seaborn\n",
      "from sklearn.tree import DecisionTreeClassifier\n",
      "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier \n",
      "from sklearn import svm\n",
      "from sklearn.metrics import confusion_matrix, accuracy_score, mean_squared_error\n",
      "from sklearn.cross_validation import cross_val_score, KFold, train_test_split\n",
      "from sklearn.decomposition import PCA\n",
      "from sklearn.preprocessing import StandardScaler\n",
      "import os.path\n",
      "\n",
      "%matplotlib inline "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Describe data ###\n",
      "\n",
      "None Variable: 2\n",
      "Quasar: 3\n",
      "Be Stars: 4\n",
      "Cepheid: 5\n",
      "RR Lyrae: 6\n",
      "Eclipsing Binaries: 7\n",
      "MicroLensing: 8\n",
      "Long Periodic Variable: 9\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "path = path = os.getcwd()\n",
      "path = path + '/MACHO_ts/'\n",
      "\n",
      "for idx,j in enumerate(os.listdir(path)):\n",
      "# for index i,j in os.listdir(path):\n",
      "    if j.endswith('.csv'):\n",
      "        if idx == 0:\n",
      "            data = pd.read_csv(path+j,index_col=0)\n",
      "            data['Class'] = idx + np.ones(len(data))\n",
      "\n",
      "        else:        \n",
      "            data2 = pd.read_csv(path+j,index_col=0)\n",
      "            data2['Class'] = idx + np.ones(len(data2))\n",
      "            data = pd.concat([data, data2])\n",
      "\n",
      "data = data.drop('Automean', 1)\n",
      "data.to_csv('MACHO_ts.csv') \n",
      "            "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# data = pd.read_csv('New_training_psi.csv',index_col=0)\n",
      "data = data.replace([np.inf, -np.inf], np.nan)\n",
      "data = data.dropna()\n",
      "data.to_csv('MACHO_ts.csv') \n",
      "data.head()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>Amplitude</th>\n",
        "      <th>AndersonDarling</th>\n",
        "      <th>Autocor_length</th>\n",
        "      <th>Beyond1Std</th>\n",
        "      <th>CAR_sigma</th>\n",
        "      <th>CAR_tau</th>\n",
        "      <th>CAR_tmean</th>\n",
        "      <th>Color</th>\n",
        "      <th>Con</th>\n",
        "      <th>Eta_color</th>\n",
        "      <th>...</th>\n",
        "      <th>Rcs</th>\n",
        "      <th>Skew</th>\n",
        "      <th>SlottedA_length</th>\n",
        "      <th>SmallKurtosis</th>\n",
        "      <th>Std</th>\n",
        "      <th>StetsonJ</th>\n",
        "      <th>StetsonK</th>\n",
        "      <th>StetsonK_AC</th>\n",
        "      <th>StetsonL</th>\n",
        "      <th>Class</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>80.6708.5218</th>\n",
        "      <td> 0.12700</td>\n",
        "      <td> 1</td>\n",
        "      <td>  44</td>\n",
        "      <td> 0.278136</td>\n",
        "      <td>-0.013622</td>\n",
        "      <td>  37.158804</td>\n",
        "      <td>-0.195121</td>\n",
        "      <td>-0.356996</td>\n",
        "      <td> 0.004307</td>\n",
        "      <td> 4478.196740</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.275917</td>\n",
        "      <td> 0.512233</td>\n",
        "      <td>   0.026157</td>\n",
        "      <td> 1.389718</td>\n",
        "      <td> 0.063573</td>\n",
        "      <td> 0.439091</td>\n",
        "      <td> 0.763764</td>\n",
        "      <td> 0.816338</td>\n",
        "      <td> 0.274887</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>48.2742.1409</th>\n",
        "      <td> 0.16525</td>\n",
        "      <td> 1</td>\n",
        "      <td>   1</td>\n",
        "      <td> 0.185185</td>\n",
        "      <td>-0.053792</td>\n",
        "      <td>   3.563116</td>\n",
        "      <td>-1.814895</td>\n",
        "      <td>-0.520699</td>\n",
        "      <td> 0.000000</td>\n",
        "      <td>   16.673054</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.207167</td>\n",
        "      <td>-1.929378</td>\n",
        "      <td>   1.024734</td>\n",
        "      <td> 7.272569</td>\n",
        "      <td> 0.076998</td>\n",
        "      <td> 0.006110</td>\n",
        "      <td> 0.673988</td>\n",
        "      <td> 0.721044</td>\n",
        "      <td> 0.003726</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>53.3970.353</th>\n",
        "      <td> 0.14600</td>\n",
        "      <td> 1</td>\n",
        "      <td>  19</td>\n",
        "      <td> 0.291367</td>\n",
        "      <td>-0.003886</td>\n",
        "      <td> 490.231454</td>\n",
        "      <td>-0.010910</td>\n",
        "      <td>-0.435427</td>\n",
        "      <td> 0.000000</td>\n",
        "      <td> 2628.279811</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.300025</td>\n",
        "      <td> 0.387355</td>\n",
        "      <td>  15.233322</td>\n",
        "      <td> 0.318671</td>\n",
        "      <td> 0.072600</td>\n",
        "      <td> 0.399034</td>\n",
        "      <td> 0.769279</td>\n",
        "      <td> 0.765331</td>\n",
        "      <td> 0.280160</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>14.8863.660</th>\n",
        "      <td> 0.13200</td>\n",
        "      <td> 1</td>\n",
        "      <td> 146</td>\n",
        "      <td> 0.316737</td>\n",
        "      <td>-0.020506</td>\n",
        "      <td>  23.280406</td>\n",
        "      <td>-0.291598</td>\n",
        "      <td>-0.260269</td>\n",
        "      <td> 0.001062</td>\n",
        "      <td> 2390.816592</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.372110</td>\n",
        "      <td>-0.547138</td>\n",
        "      <td> 306.215875</td>\n",
        "      <td>-0.031463</td>\n",
        "      <td> 0.072473</td>\n",
        "      <td> 0.721305</td>\n",
        "      <td> 0.814942</td>\n",
        "      <td> 0.743765</td>\n",
        "      <td> 0.606026</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>11.8751.968</th>\n",
        "      <td> 0.48125</td>\n",
        "      <td> 1</td>\n",
        "      <td>  41</td>\n",
        "      <td> 0.251832</td>\n",
        "      <td>-0.016820</td>\n",
        "      <td> 216.659786</td>\n",
        "      <td>-0.022690</td>\n",
        "      <td> 0.386472</td>\n",
        "      <td> 0.003670</td>\n",
        "      <td>  577.076785</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.283297</td>\n",
        "      <td> 0.464646</td>\n",
        "      <td>   0.319352</td>\n",
        "      <td> 1.554634</td>\n",
        "      <td> 0.245696</td>\n",
        "      <td> 0.599099</td>\n",
        "      <td> 0.778154</td>\n",
        "      <td> 0.718825</td>\n",
        "      <td> 0.443962</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "<p>5 rows \u00d7 65 columns</p>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 7,
       "text": [
        "              Amplitude  AndersonDarling  Autocor_length  Beyond1Std  \\\n",
        "80.6708.5218    0.12700                1              44    0.278136   \n",
        "48.2742.1409    0.16525                1               1    0.185185   \n",
        "53.3970.353     0.14600                1              19    0.291367   \n",
        "14.8863.660     0.13200                1             146    0.316737   \n",
        "11.8751.968     0.48125                1              41    0.251832   \n",
        "\n",
        "              CAR_sigma     CAR_tau  CAR_tmean     Color       Con  \\\n",
        "80.6708.5218  -0.013622   37.158804  -0.195121 -0.356996  0.004307   \n",
        "48.2742.1409  -0.053792    3.563116  -1.814895 -0.520699  0.000000   \n",
        "53.3970.353   -0.003886  490.231454  -0.010910 -0.435427  0.000000   \n",
        "14.8863.660   -0.020506   23.280406  -0.291598 -0.260269  0.001062   \n",
        "11.8751.968   -0.016820  216.659786  -0.022690  0.386472  0.003670   \n",
        "\n",
        "                Eta_color     ...            Rcs      Skew  SlottedA_length  \\\n",
        "80.6708.5218  4478.196740     ...       0.275917  0.512233         0.026157   \n",
        "48.2742.1409    16.673054     ...       0.207167 -1.929378         1.024734   \n",
        "53.3970.353   2628.279811     ...       0.300025  0.387355        15.233322   \n",
        "14.8863.660   2390.816592     ...       0.372110 -0.547138       306.215875   \n",
        "11.8751.968    577.076785     ...       0.283297  0.464646         0.319352   \n",
        "\n",
        "              SmallKurtosis       Std  StetsonJ  StetsonK  StetsonK_AC  \\\n",
        "80.6708.5218       1.389718  0.063573  0.439091  0.763764     0.816338   \n",
        "48.2742.1409       7.272569  0.076998  0.006110  0.673988     0.721044   \n",
        "53.3970.353        0.318671  0.072600  0.399034  0.769279     0.765331   \n",
        "14.8863.660       -0.031463  0.072473  0.721305  0.814942     0.743765   \n",
        "11.8751.968        1.554634  0.245696  0.599099  0.778154     0.718825   \n",
        "\n",
        "              StetsonL  Class  \n",
        "80.6708.5218  0.274887      1  \n",
        "48.2742.1409  0.003726      1  \n",
        "53.3970.353   0.280160      1  \n",
        "14.8863.660   0.606026      1  \n",
        "11.8751.968   0.443962      1  \n",
        "\n",
        "[5 rows x 65 columns]"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "features = data.columns[:-1]\n",
      "data[features].describe()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>Amplitude</th>\n",
        "      <th>AndersonDarling</th>\n",
        "      <th>Autocor_length</th>\n",
        "      <th>Beyond1Std</th>\n",
        "      <th>CAR_sigma</th>\n",
        "      <th>CAR_tau</th>\n",
        "      <th>CAR_tmean</th>\n",
        "      <th>Color</th>\n",
        "      <th>Con</th>\n",
        "      <th>Eta_color</th>\n",
        "      <th>...</th>\n",
        "      <th>Q31_color</th>\n",
        "      <th>Rcs</th>\n",
        "      <th>Skew</th>\n",
        "      <th>SlottedA_length</th>\n",
        "      <th>SmallKurtosis</th>\n",
        "      <th>Std</th>\n",
        "      <th>StetsonJ</th>\n",
        "      <th>StetsonK</th>\n",
        "      <th>StetsonK_AC</th>\n",
        "      <th>StetsonL</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>count</th>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6.063000e+03</td>\n",
        "      <td> 6.063000e+03</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td>    6063.000000</td>\n",
        "      <td>...</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "      <td> 6063.000000</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>mean</th>\n",
        "      <td>    0.523354</td>\n",
        "      <td>    0.998227</td>\n",
        "      <td>    4.366320</td>\n",
        "      <td>    0.240932</td>\n",
        "      <td>   -0.078952</td>\n",
        "      <td> 6.221701e+05</td>\n",
        "      <td>-2.275857e+02</td>\n",
        "      <td>    0.204879</td>\n",
        "      <td>    0.002929</td>\n",
        "      <td>    7494.105158</td>\n",
        "      <td>...</td>\n",
        "      <td>    0.258005</td>\n",
        "      <td>    0.070918</td>\n",
        "      <td>   -0.364092</td>\n",
        "      <td>    3.885572</td>\n",
        "      <td>    4.659390</td>\n",
        "      <td>    0.253619</td>\n",
        "      <td>    0.258613</td>\n",
        "      <td>    0.709223</td>\n",
        "      <td>    0.676494</td>\n",
        "      <td>    0.165221</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>std</th>\n",
        "      <td>    0.335874</td>\n",
        "      <td>    0.025019</td>\n",
        "      <td>   17.387283</td>\n",
        "      <td>    0.069362</td>\n",
        "      <td>    1.392112</td>\n",
        "      <td> 2.277031e+07</td>\n",
        "      <td> 8.347666e+02</td>\n",
        "      <td>    0.528046</td>\n",
        "      <td>    0.008102</td>\n",
        "      <td>   42197.710030</td>\n",
        "      <td>...</td>\n",
        "      <td>    0.184689</td>\n",
        "      <td>    0.062818</td>\n",
        "      <td>    0.794886</td>\n",
        "      <td>   28.955520</td>\n",
        "      <td>    4.733775</td>\n",
        "      <td>    0.161772</td>\n",
        "      <td>    0.246856</td>\n",
        "      <td>    0.067021</td>\n",
        "      <td>    0.094042</td>\n",
        "      <td>    0.201636</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>min</th>\n",
        "      <td>    0.020500</td>\n",
        "      <td>    0.341722</td>\n",
        "      <td>    1.000000</td>\n",
        "      <td>    0.044944</td>\n",
        "      <td>  -12.148201</td>\n",
        "      <td> 4.728889e-04</td>\n",
        "      <td>-1.091337e+04</td>\n",
        "      <td>   -1.022921</td>\n",
        "      <td>    0.000000</td>\n",
        "      <td>       1.000000</td>\n",
        "      <td>...</td>\n",
        "      <td>    0.009000</td>\n",
        "      <td>    0.019402</td>\n",
        "      <td>   -6.136411</td>\n",
        "      <td>    0.011111</td>\n",
        "      <td>   -1.515619</td>\n",
        "      <td>    0.011412</td>\n",
        "      <td>   -0.177488</td>\n",
        "      <td>    0.395244</td>\n",
        "      <td>    0.250155</td>\n",
        "      <td>   -0.114115</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>25%</th>\n",
        "      <td>    0.261125</td>\n",
        "      <td>    1.000000</td>\n",
        "      <td>    1.000000</td>\n",
        "      <td>    0.197250</td>\n",
        "      <td>   -0.217596</td>\n",
        "      <td> 4.820693e-02</td>\n",
        "      <td>-1.023507e+02</td>\n",
        "      <td>   -0.164665</td>\n",
        "      <td>    0.000000</td>\n",
        "      <td>    1391.829363</td>\n",
        "      <td>...</td>\n",
        "      <td>    0.114000</td>\n",
        "      <td>    0.038937</td>\n",
        "      <td>   -0.739871</td>\n",
        "      <td>    0.109699</td>\n",
        "      <td>    2.082438</td>\n",
        "      <td>    0.129247</td>\n",
        "      <td>    0.076718</td>\n",
        "      <td>    0.664716</td>\n",
        "      <td>    0.620217</td>\n",
        "      <td>    0.035334</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>50%</th>\n",
        "      <td>    0.465500</td>\n",
        "      <td>    1.000000</td>\n",
        "      <td>    1.000000</td>\n",
        "      <td>    0.222101</td>\n",
        "      <td>   -0.028737</td>\n",
        "      <td> 1.160218e+00</td>\n",
        "      <td>-3.728635e+00</td>\n",
        "      <td>    0.091812</td>\n",
        "      <td>    0.000000</td>\n",
        "      <td>    3713.568491</td>\n",
        "      <td>...</td>\n",
        "      <td>    0.210000</td>\n",
        "      <td>    0.049840</td>\n",
        "      <td>   -0.419509</td>\n",
        "      <td>    0.123739</td>\n",
        "      <td>    4.212347</td>\n",
        "      <td>    0.228767</td>\n",
        "      <td>    0.144524</td>\n",
        "      <td>    0.694725</td>\n",
        "      <td>    0.663081</td>\n",
        "      <td>    0.067666</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>75%</th>\n",
        "      <td>    0.721750</td>\n",
        "      <td>    1.000000</td>\n",
        "      <td>    1.000000</td>\n",
        "      <td>    0.263378</td>\n",
        "      <td>    0.018142</td>\n",
        "      <td> 1.832332e+01</td>\n",
        "      <td>-2.662364e-01</td>\n",
        "      <td>    0.406176</td>\n",
        "      <td>    0.001701</td>\n",
        "      <td>    7592.668454</td>\n",
        "      <td>...</td>\n",
        "      <td>    0.369000</td>\n",
        "      <td>    0.070661</td>\n",
        "      <td>   -0.008725</td>\n",
        "      <td>    0.203970</td>\n",
        "      <td>    6.440439</td>\n",
        "      <td>    0.346826</td>\n",
        "      <td>    0.416958</td>\n",
        "      <td>    0.742183</td>\n",
        "      <td>    0.726495</td>\n",
        "      <td>    0.214128</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>max</th>\n",
        "      <td>    3.221250</td>\n",
        "      <td>    1.000000</td>\n",
        "      <td>  323.000000</td>\n",
        "      <td>    0.788321</td>\n",
        "      <td>   25.646470</td>\n",
        "      <td> 1.379351e+09</td>\n",
        "      <td>-3.597353e-09</td>\n",
        "      <td>    3.263769</td>\n",
        "      <td>    0.143713</td>\n",
        "      <td> 1807743.974233</td>\n",
        "      <td>...</td>\n",
        "      <td>    1.469000</td>\n",
        "      <td>    0.462523</td>\n",
        "      <td>   10.406876</td>\n",
        "      <td>  527.874240</td>\n",
        "      <td>  172.925347</td>\n",
        "      <td>    1.718985</td>\n",
        "      <td>    1.414214</td>\n",
        "      <td>    0.928697</td>\n",
        "      <td>    0.952755</td>\n",
        "      <td>    1.772197</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "<p>8 rows \u00d7 64 columns</p>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 8,
       "text": [
        "         Amplitude  AndersonDarling  Autocor_length   Beyond1Std    CAR_sigma  \\\n",
        "count  6063.000000      6063.000000     6063.000000  6063.000000  6063.000000   \n",
        "mean      0.523354         0.998227        4.366320     0.240932    -0.078952   \n",
        "std       0.335874         0.025019       17.387283     0.069362     1.392112   \n",
        "min       0.020500         0.341722        1.000000     0.044944   -12.148201   \n",
        "25%       0.261125         1.000000        1.000000     0.197250    -0.217596   \n",
        "50%       0.465500         1.000000        1.000000     0.222101    -0.028737   \n",
        "75%       0.721750         1.000000        1.000000     0.263378     0.018142   \n",
        "max       3.221250         1.000000      323.000000     0.788321    25.646470   \n",
        "\n",
        "            CAR_tau     CAR_tmean        Color          Con       Eta_color  \\\n",
        "count  6.063000e+03  6.063000e+03  6063.000000  6063.000000     6063.000000   \n",
        "mean   6.221701e+05 -2.275857e+02     0.204879     0.002929     7494.105158   \n",
        "std    2.277031e+07  8.347666e+02     0.528046     0.008102    42197.710030   \n",
        "min    4.728889e-04 -1.091337e+04    -1.022921     0.000000        1.000000   \n",
        "25%    4.820693e-02 -1.023507e+02    -0.164665     0.000000     1391.829363   \n",
        "50%    1.160218e+00 -3.728635e+00     0.091812     0.000000     3713.568491   \n",
        "75%    1.832332e+01 -2.662364e-01     0.406176     0.001701     7592.668454   \n",
        "max    1.379351e+09 -3.597353e-09     3.263769     0.143713  1807743.974233   \n",
        "\n",
        "            ...          Q31_color          Rcs         Skew  SlottedA_length  \\\n",
        "count       ...        6063.000000  6063.000000  6063.000000      6063.000000   \n",
        "mean        ...           0.258005     0.070918    -0.364092         3.885572   \n",
        "std         ...           0.184689     0.062818     0.794886        28.955520   \n",
        "min         ...           0.009000     0.019402    -6.136411         0.011111   \n",
        "25%         ...           0.114000     0.038937    -0.739871         0.109699   \n",
        "50%         ...           0.210000     0.049840    -0.419509         0.123739   \n",
        "75%         ...           0.369000     0.070661    -0.008725         0.203970   \n",
        "max         ...           1.469000     0.462523    10.406876       527.874240   \n",
        "\n",
        "       SmallKurtosis          Std     StetsonJ     StetsonK  StetsonK_AC  \\\n",
        "count    6063.000000  6063.000000  6063.000000  6063.000000  6063.000000   \n",
        "mean        4.659390     0.253619     0.258613     0.709223     0.676494   \n",
        "std         4.733775     0.161772     0.246856     0.067021     0.094042   \n",
        "min        -1.515619     0.011412    -0.177488     0.395244     0.250155   \n",
        "25%         2.082438     0.129247     0.076718     0.664716     0.620217   \n",
        "50%         4.212347     0.228767     0.144524     0.694725     0.663081   \n",
        "75%         6.440439     0.346826     0.416958     0.742183     0.726495   \n",
        "max       172.925347     1.718985     1.414214     0.928697     0.952755   \n",
        "\n",
        "          StetsonL  \n",
        "count  6063.000000  \n",
        "mean      0.165221  \n",
        "std       0.201636  \n",
        "min      -0.114115  \n",
        "25%       0.035334  \n",
        "50%       0.067666  \n",
        "75%       0.214128  \n",
        "max       1.772197  \n",
        "\n",
        "[8 rows x 64 columns]"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print 'Data size:',len(data)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Data size: 6063\n"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print \"number of features: %i \" %(len(data.columns[:-1]))\n",
      "\n",
      "counts = pd.DataFrame(data.dropna().groupby(by='Class').CAR_sigma.count())\n",
      "counts.columns = ['count']\n",
      "counts"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "number of features: 64 \n"
       ]
      },
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>count</th>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>Class</th>\n",
        "      <th></th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>1</th>\n",
        "      <td>  127</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>2</th>\n",
        "      <td>  101</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>3</th>\n",
        "      <td>  255</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>4</th>\n",
        "      <td>  365</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>5</th>\n",
        "      <td>  580</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>6</th>\n",
        "      <td> 3966</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>7</th>\n",
        "      <td>   59</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>8</th>\n",
        "      <td>  610</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 6,
       "text": [
        "       count\n",
        "Class       \n",
        "1        127\n",
        "2        101\n",
        "3        255\n",
        "4        365\n",
        "5        580\n",
        "6       3966\n",
        "7         59\n",
        "8        610"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "y = data['Class']\n",
      "plt.scatter(data.iloc[:,4], data.iloc[:,10], c=y, s=50)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 98,
       "text": [
        "<matplotlib.collections.PathCollection at 0x10ce96450>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEACAYAAABPiSrXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nv1X0AAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10ce2df10>"
       ]
      }
     ],
     "prompt_number": 98
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Random forest classification ###"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "Xr, Xt, yr, yt = train_test_split(data[features], data['Class'], train_size=5000, test_size=1063, random_state=42)\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 99
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf = RandomForestClassifier(n_estimators=300)\n",
      "clf.fit(Xr, yr)\n",
      "\n",
      "ypred = clf.predict(Xt)\n",
      "print 'confusion matrix: \\n' \n",
      "print (confusion_matrix(yt, ypred))\n",
      "print '\\n ========================================= \\n '\n",
      "print 'score of accuracy %f' % (clf.score(Xt, yt))\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "confusion matrix: \n",
        "\n",
        "[[ 20   0   0   1   1   1   3   1]\n",
        " [  0  17   1   0   0   0   0   0]\n",
        " [  0   0  49   1   1   2   0   6]\n",
        " [  0   0   0  60   2   0   0   0]\n",
        " [  0   0   0   0 100   2   1   0]\n",
        " [  0   1   1   1   0 685   0   1]\n",
        " [  2   0   0   0   0   0   8   0]\n",
        " [  0   0   0   0   0   0   0  95]]\n",
        "\n",
        " ========================================= \n",
        " \n",
        "score of accuracy 0.972719"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n"
       ]
      }
     ],
     "prompt_number": 100
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "scores = cross_val_score(clf, Xt, yt, cv=2)\n",
      "print(\"Accuracy: %0.2f (+/- %0.2f)\"\n",
      "      % (scores.mean(), scores.std()*2))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Accuracy: 0.96 (+/- 0.00)\n"
       ]
      }
     ],
     "prompt_number": 101
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def plot_estimator(estimator, X, y):\n",
      "    estimator.fit(X, y)\n",
      "    x_min, x_max = X[:, 0].min() - .1, X[:, 0].max() + .1\n",
      "    y_min, y_max = X[:, 1].min() - .1, X[:, 1].max() + .1\n",
      "    xx, yy = np.meshgrid(np.linspace(x_min, x_max, 50),\n",
      "                         np.linspace(y_min, y_max, 50))\n",
      "    Z = estimator.predict(np.c_[xx.ravel(), yy.ravel()])\n",
      "\n",
      "    # Put the result into a color plot\n",
      "    Z = Z.reshape(xx.shape)\n",
      "    plt.figure(figsize=[8,10])\n",
      "    plt.pcolormesh(xx, yy, Z, alpha=0.15,cmap = 'hsv')\n",
      "\n",
      "    # Plot also the training points\n",
      "    plt.scatter(X[:, 0], X[:, 1], c = y, s=50)\n",
      "    plt.axis('tight')\n",
      "    plt.axis('off')\n",
      "    plt.tight_layout()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 90
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plot_estimator(clf, data[features], data['Class'].values)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "ename": "TypeError",
       "evalue": "unhashable type",
       "output_type": "pyerr",
       "traceback": [
        "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m\n\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
        "\u001b[0;32m<ipython-input-91-21150e44bc76>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mplot_estimator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclf\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mfeatures\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'Class'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
        "\u001b[0;32m<ipython-input-90-9a3a7126659c>\u001b[0m in \u001b[0;36mplot_estimator\u001b[0;34m(estimator, X, y)\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mplot_estimator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mestimator\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m     \u001b[0mestimator\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m     \u001b[0mx_min\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx_max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mX\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m.1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m.1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m     \u001b[0my_min\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mX\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m.1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m.1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     xx, yy = np.meshgrid(np.linspace(x_min, x_max, 50),\n",
        "\u001b[0;32m/Users/isadoranun/anaconda/lib/python2.7/site-packages/pandas/core/frame.pyc\u001b[0m in \u001b[0;36m__getitem__\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m   1676\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_getitem_multilevel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1677\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1678\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_getitem_column\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1679\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1680\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_getitem_column\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;32m/Users/isadoranun/anaconda/lib/python2.7/site-packages/pandas/core/frame.pyc\u001b[0m in \u001b[0;36m_getitem_column\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m   1683\u001b[0m         \u001b[0;31m# get column\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1684\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcolumns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_unique\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1685\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_get_item_cache\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1686\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1687\u001b[0m         \u001b[0;31m# duplicate columns & possible reduce dimensionaility\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;32m/Users/isadoranun/anaconda/lib/python2.7/site-packages/pandas/core/generic.pyc\u001b[0m in \u001b[0;36m_get_item_cache\u001b[0;34m(self, item)\u001b[0m\n\u001b[1;32m   1048\u001b[0m         \u001b[0;34m\"\"\" return the cached item, item represents a label indexer \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1049\u001b[0m         \u001b[0mcache\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_item_cache\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1050\u001b[0;31m         \u001b[0mres\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcache\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mitem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1051\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mres\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1052\u001b[0m             \u001b[0mvalues\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_data\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mitem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;31mTypeError\u001b[0m: unhashable type"
       ]
      }
     ],
     "prompt_number": 91
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### pruning for $ \\textit{max_depths} $ for decision tree ###"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "\n",
      "kfold = KFold(Xr.shape[0], n_folds=10)\n",
      "accs = []\n",
      "max_depths = range(1, 20)\n",
      "for max_depth in max_depths:\n",
      "    k_accs = []\n",
      "    for train, test in kfold:\n",
      "        Xtrain, Xtest, ytrain, ytest = Xr[train], Xr[test], yr[train], yr[test]\n",
      "        clf = DecisionTreeClassifier(max_depth=max_depth)\n",
      "        clf.fit(Xtrain, ytrain)\n",
      "        ypred = clf.predict(Xtest)\n",
      "        k_accs.append(accuracy_score(ytest, ypred))\n",
      "    accs.append(np.mean(k_accs))\n",
      "# plot the accuracies as a function of max_depth\n",
      "plt.plot(max_depths, accs, linewidth=2.5)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 72,
       "text": [
        "[<matplotlib.lines.Line2D at 0x10be0bb50>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAXsAAAEACAYAAABS29YJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncd70QBQAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10b407310>"
       ]
      }
     ],
     "prompt_number": 72
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Boosting to tune $ \\textit{number of trees} $ ###\n",
      "\n",
      "Boosting builds lots of small trees. Unlike Random Forests, each tree tries to make up for the deficiencies in the current ensemble. With Boosting, the number of trees is a tuning parameter. Using the same train/test split, we plot the variance of MSE against the number of trees."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "num_trees = range(100, 2000, 100)\n",
      "mses = []\n",
      "for num_tree in num_trees:\n",
      "    print (num_tree)\n",
      "    reg = RandomForestClassifier(n_estimators=num_tree)\n",
      "    reg.fit(Xr, yr)\n",
      "    ypred = reg.predict(Xt)\n",
      "    mses.append(mean_squared_error(yt, ypred))\n",
      "plt.plot(num_trees, mses)\n",
      "plt.xlabel(\"#-trees\")\n",
      "plt.ylabel(\"MSE\")"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "100\n",
        "200"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "300"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "400"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "500"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "600"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "700"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "800"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "900"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1000"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1100"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1200"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1300"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1400"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1500"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1600"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1700"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1800"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "1900"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n"
       ]
      },
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 73,
       "text": [
        "<matplotlib.text.Text at 0x10b2b02d0>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAZEAAAEPCAYAAACDTflkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10a5b2c90>"
       ]
      }
     ],
     "prompt_number": 73
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### tuning for $ \\textit{max_features} $ ###"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "oob_scores = []\n",
      "mses = []\n",
      "num_feats = range(1, 33)\n",
      "for num_feat in num_feats:\n",
      "    clf = RandomForestClassifier(n_estimators=500, max_features=num_feat, oob_score=True)\n",
      "    clf.fit(Xr, yr)\n",
      "    ypred = clf.predict(Xt)\n",
      "    mses.append(mean_squared_error(yt, ypred))\n",
      "    oob_scores.append(clf.oob_score_)\n",
      "plt.plot(num_feats, mses, color='b', linewidth=2.5);\n",
      "plt.plot(num_feats, oob_scores, color='r', linewidth=2.5);\n",
      "plt.xlabel(\"features\")\n",
      "plt.ylabel(\"MSE (blue); OOB (red)\")"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 74,
       "text": [
        "<matplotlib.text.Text at 0x10f7c0f90>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEPCAYAAABCyrPIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\noIQgIiIRJQQREQHg/wPwe0ih1duUKAAAAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x11043c6d0>"
       ]
      }
     ],
     "prompt_number": 74
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### standard scale data to center around 0 ###"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def stand_features(data, feature_names = 'all'):\n",
      "    if feature_names == 'all': \n",
      "        column_names = data.columns\n",
      "        data = pd.DataFrame(StandardScaler().fit_transform(data), columns=column_names)\n",
      "    else:\n",
      "        data[feature_names] = pd.DataFrame(StandardScaler().fit_transform(data), columns=feature_names)\n",
      "    return data       "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 56
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "newData = stand_features(train)\n",
      "newData.head()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "ename": "AttributeError",
       "evalue": "'numpy.ndarray' object has no attribute 'columns'",
       "output_type": "pyerr",
       "traceback": [
        "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m\n\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
        "\u001b[0;32m<ipython-input-57-f299c776d488>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mnewData\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstand_features\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mnewData\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhead\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;32m<ipython-input-56-9f3c4a89cb56>\u001b[0m in \u001b[0;36mstand_features\u001b[0;34m(data, feature_names)\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mstand_features\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeature_names\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m'all'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mfeature_names\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'all'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m         \u001b[0mcolumn_names\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcolumns\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m         \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mDataFrame\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mStandardScaler\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit_transform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcolumns\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcolumn_names\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;31mAttributeError\u001b[0m: 'numpy.ndarray' object has no attribute 'columns'"
       ]
      }
     ],
     "prompt_number": 57
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf = RandomForestClassifier(n_estimators=100)\n",
      "clf.fit(Xr, yr)\n",
      "\n",
      "print clf.score(Xt, yt)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "0.962\n"
       ]
      }
     ],
     "prompt_number": 58
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### random forest for feature importance ###"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "** Features used at the top of the tree are used contribute to the final prediction decision of a larger fraction of the input samples. The expected fraction of the samples they contribute to can thus be used as an estimate of the relative importance of the features. **"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# look up the specific algorithm\n",
      "importances = clf.feature_importances_\n",
      "sorted_idx = np.argsort(importances)\n",
      "\n",
      "plt.figure(figsize = [10,12])\n",
      "padding = np.arange(len(features)) + 0.2\n",
      "plt.barh(padding, importances[sorted_idx], align='center')\n",
      "plt.yticks(padding, features[sorted_idx])\n",
      "plt.xlabel(\"Relative Importance\")\n",
      "plt.title(\"Variable Importance\")\n",
      "plt.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAukAAALYCAYAAADB4mwDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZmZmZiXjTrqZmZmZWcn8/ymQk6B6fuKUAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10cda8e50>"
       ]
      }
     ],
     "prompt_number": 102
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### SVM for classification ###\n",
      "\n",
      "    1. Linear Kernal: no real transformation on the data\n",
      "        Parameters: C (the ratio of how much emphasis we want to put on minimizing the error or maximizing the margin. Large C: Lower bias, higher variance, pro to overfitting)\n",
      "    2. Polynomial Kernal: of degre k on n original features yields O(n^k) expanded features\n",
      "        Parameters: C, degree(of the polynomial), gamma(coefficient of the polynomial)\n",
      "    3. RBF Kernal: normal, yield infinitely many expanded features\n",
      "        Parameters: gamma(1/gamma is the variance)"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf = svm.linearSVC()\n",
      "clf.fit(Xr, yr)\n",
      "print 'score of accuracy %f' % (clf.score(Xt, yt))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "ename": "AttributeError",
       "evalue": "'module' object has no attribute 'linearSVC'",
       "output_type": "pyerr",
       "traceback": [
        "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m\n\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
        "\u001b[0;32m<ipython-input-48-02d854a67da1>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mclf\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msvm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlinearSVC\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mclf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;32mprint\u001b[0m \u001b[0;34m'score of accuracy %f'\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mclf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mscore\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;31mAttributeError\u001b[0m: 'module' object has no attribute 'linearSVC'"
       ]
      }
     ],
     "prompt_number": 48
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### deep learning ###"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import theano\n",
      "import theano.tensor as T"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "theano.__version__\n",
      "print theano.config.device"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "cpu\n"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "pd.reset_option('display.max_rows')\n",
      "\n",
      "path ='/Users/isadoranun/Desktop/Features/F_1/'\n",
      "\n",
      "\n",
      "data1 = pd.read_csv('F_1_3319.csv',index_col=0)\n",
      "data2 = pd.read_csv('F_1_3319_2.csv',index_col=0)\n",
      "\n",
      "data1.Rcs = data2.Rcs\n",
      "\n",
      "features = data2.columns[1:-1]\n",
      "data2 = data2[features]\n",
      "\n",
      "concatenate = pd.merge(data1,data2, left_index=True, right_index=True, how='outer')\n",
      "concatenate.head()\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>Amplitude</th>\n",
        "      <th>AndersonDarling</th>\n",
        "      <th>B_R</th>\n",
        "      <th>Beyond1Std</th>\n",
        "      <th>Bmean</th>\n",
        "      <th>CAR_sigma</th>\n",
        "      <th>CAR_tau</th>\n",
        "      <th>CAR_tmean</th>\n",
        "      <th>Con</th>\n",
        "      <th>Eta_B_R</th>\n",
        "      <th>...</th>\n",
        "      <th>Std</th>\n",
        "      <th>StetsonJ</th>\n",
        "      <th>StetsonK</th>\n",
        "      <th>StetsonK_AC</th>\n",
        "      <th>StetsonL</th>\n",
        "      <th>autocor</th>\n",
        "      <th>automean</th>\n",
        "      <th>meanvariance</th>\n",
        "      <th>Psi_CS</th>\n",
        "      <th>Psi_eta</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>1.3319.23</th>\n",
        "      <td> 0.42450</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0.249543</td>\n",
        "      <td> 0.196115</td>\n",
        "      <td>-4.938551</td>\n",
        "      <td>-0.165672</td>\n",
        "      <td>   1.206547</td>\n",
        "      <td>-4.093127</td>\n",
        "      <td> 0.000000</td>\n",
        "      <td> 6818.615959</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.199041</td>\n",
        "      <td> 0.187266</td>\n",
        "      <td> 0.637712</td>\n",
        "      <td> 0.457122</td>\n",
        "      <td> 0.100231</td>\n",
        "      <td>  2</td>\n",
        "      <td>-4.938551</td>\n",
        "      <td>-0.040304</td>\n",
        "      <td> 0.085552</td>\n",
        "      <td> 1.836131</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1.3319.11</th>\n",
        "      <td> 0.06800</td>\n",
        "      <td> 1</td>\n",
        "      <td>-0.409863</td>\n",
        "      <td> 0.264388</td>\n",
        "      <td>-7.432300</td>\n",
        "      <td>-0.028999</td>\n",
        "      <td>   3.137702</td>\n",
        "      <td>-2.368708</td>\n",
        "      <td> 0.004505</td>\n",
        "      <td> 3426.229184</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.035677</td>\n",
        "      <td> 0.435276</td>\n",
        "      <td> 0.716623</td>\n",
        "      <td> 0.875010</td>\n",
        "      <td> 0.196524</td>\n",
        "      <td> 13</td>\n",
        "      <td>-7.432300</td>\n",
        "      <td>-0.004800</td>\n",
        "      <td> 0.151780</td>\n",
        "      <td> 1.644436</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1.3319.42</th>\n",
        "      <td> 0.73575</td>\n",
        "      <td> 1</td>\n",
        "      <td>-0.278173</td>\n",
        "      <td> 0.229023</td>\n",
        "      <td>-4.117237</td>\n",
        "      <td> 0.226868</td>\n",
        "      <td>   2.019573</td>\n",
        "      <td>-2.038667</td>\n",
        "      <td> 0.000000</td>\n",
        "      <td> 2511.253694</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.360103</td>\n",
        "      <td> 0.213843</td>\n",
        "      <td> 0.707300</td>\n",
        "      <td> 0.482501</td>\n",
        "      <td> 0.136990</td>\n",
        "      <td>  2</td>\n",
        "      <td>-4.117237</td>\n",
        "      <td>-0.087462</td>\n",
        "      <td> 0.059685</td>\n",
        "      <td> 1.871534</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1.3319.33</th>\n",
        "      <td> 0.84100</td>\n",
        "      <td> 1</td>\n",
        "      <td>-0.281362</td>\n",
        "      <td> 0.232558</td>\n",
        "      <td>-4.056352</td>\n",
        "      <td> 0.019112</td>\n",
        "      <td> 251.327957</td>\n",
        "      <td>-0.016140</td>\n",
        "      <td> 0.000000</td>\n",
        "      <td>  632.585112</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.440170</td>\n",
        "      <td> 0.325442</td>\n",
        "      <td> 0.727188</td>\n",
        "      <td> 0.464172</td>\n",
        "      <td> 0.213371</td>\n",
        "      <td>  2</td>\n",
        "      <td>-4.056352</td>\n",
        "      <td>-0.108514</td>\n",
        "      <td> 0.058776</td>\n",
        "      <td> 1.958520</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1.3319.34</th>\n",
        "      <td> 0.35600</td>\n",
        "      <td> 1</td>\n",
        "      <td>-0.720231</td>\n",
        "      <td> 0.205993</td>\n",
        "      <td>-4.844879</td>\n",
        "      <td>-0.154069</td>\n",
        "      <td>   0.974836</td>\n",
        "      <td>-4.969945</td>\n",
        "      <td> 0.000938</td>\n",
        "      <td> 5246.245871</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.179034</td>\n",
        "      <td> 0.181015</td>\n",
        "      <td> 0.674830</td>\n",
        "      <td> 0.407155</td>\n",
        "      <td> 0.086707</td>\n",
        "      <td>  2</td>\n",
        "      <td>-4.844879</td>\n",
        "      <td>-0.036953</td>\n",
        "      <td> 0.047025</td>\n",
        "      <td> 1.992605</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "<p>5 rows \u00d7 41 columns</p>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 113,
       "text": [
        "           Amplitude  AndersonDarling       B_R  Beyond1Std     Bmean  \\\n",
        "1.3319.23    0.42450                1  0.249543    0.196115 -4.938551   \n",
        "1.3319.11    0.06800                1 -0.409863    0.264388 -7.432300   \n",
        "1.3319.42    0.73575                1 -0.278173    0.229023 -4.117237   \n",
        "1.3319.33    0.84100                1 -0.281362    0.232558 -4.056352   \n",
        "1.3319.34    0.35600                1 -0.720231    0.205993 -4.844879   \n",
        "\n",
        "           CAR_sigma     CAR_tau  CAR_tmean       Con      Eta_B_R  \\\n",
        "1.3319.23  -0.165672    1.206547  -4.093127  0.000000  6818.615959   \n",
        "1.3319.11  -0.028999    3.137702  -2.368708  0.004505  3426.229184   \n",
        "1.3319.42   0.226868    2.019573  -2.038667  0.000000  2511.253694   \n",
        "1.3319.33   0.019112  251.327957  -0.016140  0.000000   632.585112   \n",
        "1.3319.34  -0.154069    0.974836  -4.969945  0.000938  5246.245871   \n",
        "\n",
        "              ...            Std  StetsonJ  StetsonK  StetsonK_AC  StetsonL  \\\n",
        "1.3319.23     ...       0.199041  0.187266  0.637712     0.457122  0.100231   \n",
        "1.3319.11     ...       0.035677  0.435276  0.716623     0.875010  0.196524   \n",
        "1.3319.42     ...       0.360103  0.213843  0.707300     0.482501  0.136990   \n",
        "1.3319.33     ...       0.440170  0.325442  0.727188     0.464172  0.213371   \n",
        "1.3319.34     ...       0.179034  0.181015  0.674830     0.407155  0.086707   \n",
        "\n",
        "           autocor  automean  meanvariance    Psi_CS   Psi_eta  \n",
        "1.3319.23        2 -4.938551     -0.040304  0.085552  1.836131  \n",
        "1.3319.11       13 -7.432300     -0.004800  0.151780  1.644436  \n",
        "1.3319.42        2 -4.117237     -0.087462  0.059685  1.871534  \n",
        "1.3319.33        2 -4.056352     -0.108514  0.058776  1.958520  \n",
        "1.3319.34        2 -4.844879     -0.036953  0.047025  1.992605  \n",
        "\n",
        "[5 rows x 41 columns]"
       ]
      }
     ],
     "prompt_number": 113
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "data.to_csv('yelp_academic_dataset_user.csv') "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 10
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "data.head()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>Amplitude</th>\n",
        "      <th>AndersonDarling</th>\n",
        "      <th>Autocor_length</th>\n",
        "      <th>Beyond1Std</th>\n",
        "      <th>CAR_sigma</th>\n",
        "      <th>CAR_tau</th>\n",
        "      <th>CAR_tmean</th>\n",
        "      <th>Color</th>\n",
        "      <th>Con</th>\n",
        "      <th>Eta_color</th>\n",
        "      <th>...</th>\n",
        "      <th>Rcs</th>\n",
        "      <th>Skew</th>\n",
        "      <th>SlottedA_length</th>\n",
        "      <th>SmallKurtosis</th>\n",
        "      <th>Std</th>\n",
        "      <th>StetsonJ</th>\n",
        "      <th>StetsonK</th>\n",
        "      <th>StetsonK_AC</th>\n",
        "      <th>StetsonL</th>\n",
        "      <th>Class</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>80.6708.5218</th>\n",
        "      <td> 0.12700</td>\n",
        "      <td> 1</td>\n",
        "      <td>  44</td>\n",
        "      <td> 0.278136</td>\n",
        "      <td>-0.013622</td>\n",
        "      <td>  37.158804</td>\n",
        "      <td>-0.195121</td>\n",
        "      <td>-0.356996</td>\n",
        "      <td> 0.004307</td>\n",
        "      <td> 4478.196740</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.275917</td>\n",
        "      <td> 0.512233</td>\n",
        "      <td>   0.026157</td>\n",
        "      <td> 1.389718</td>\n",
        "      <td> 0.063573</td>\n",
        "      <td> 0.439091</td>\n",
        "      <td> 0.763764</td>\n",
        "      <td> 0.816338</td>\n",
        "      <td> 0.274887</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>48.2742.1409</th>\n",
        "      <td> 0.16525</td>\n",
        "      <td> 1</td>\n",
        "      <td>   1</td>\n",
        "      <td> 0.185185</td>\n",
        "      <td>-0.053792</td>\n",
        "      <td>   3.563116</td>\n",
        "      <td>-1.814895</td>\n",
        "      <td>-0.520699</td>\n",
        "      <td> 0.000000</td>\n",
        "      <td>   16.673054</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.207167</td>\n",
        "      <td>-1.929378</td>\n",
        "      <td>   1.024734</td>\n",
        "      <td> 7.272569</td>\n",
        "      <td> 0.076998</td>\n",
        "      <td> 0.006110</td>\n",
        "      <td> 0.673988</td>\n",
        "      <td> 0.721044</td>\n",
        "      <td> 0.003726</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>53.3970.353</th>\n",
        "      <td> 0.14600</td>\n",
        "      <td> 1</td>\n",
        "      <td>  19</td>\n",
        "      <td> 0.291367</td>\n",
        "      <td>-0.003886</td>\n",
        "      <td> 490.231454</td>\n",
        "      <td>-0.010910</td>\n",
        "      <td>-0.435427</td>\n",
        "      <td> 0.000000</td>\n",
        "      <td> 2628.279811</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.300025</td>\n",
        "      <td> 0.387355</td>\n",
        "      <td>  15.233322</td>\n",
        "      <td> 0.318671</td>\n",
        "      <td> 0.072600</td>\n",
        "      <td> 0.399034</td>\n",
        "      <td> 0.769279</td>\n",
        "      <td> 0.765331</td>\n",
        "      <td> 0.280160</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>14.8863.660</th>\n",
        "      <td> 0.13200</td>\n",
        "      <td> 1</td>\n",
        "      <td> 146</td>\n",
        "      <td> 0.316737</td>\n",
        "      <td>-0.020506</td>\n",
        "      <td>  23.280406</td>\n",
        "      <td>-0.291598</td>\n",
        "      <td>-0.260269</td>\n",
        "      <td> 0.001062</td>\n",
        "      <td> 2390.816592</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.372110</td>\n",
        "      <td>-0.547138</td>\n",
        "      <td> 306.215875</td>\n",
        "      <td>-0.031463</td>\n",
        "      <td> 0.072473</td>\n",
        "      <td> 0.721305</td>\n",
        "      <td> 0.814942</td>\n",
        "      <td> 0.743765</td>\n",
        "      <td> 0.606026</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>11.8751.968</th>\n",
        "      <td> 0.48125</td>\n",
        "      <td> 1</td>\n",
        "      <td>  41</td>\n",
        "      <td> 0.251832</td>\n",
        "      <td>-0.016820</td>\n",
        "      <td> 216.659786</td>\n",
        "      <td>-0.022690</td>\n",
        "      <td> 0.386472</td>\n",
        "      <td> 0.003670</td>\n",
        "      <td>  577.076785</td>\n",
        "      <td>...</td>\n",
        "      <td> 0.283297</td>\n",
        "      <td> 0.464646</td>\n",
        "      <td>   0.319352</td>\n",
        "      <td> 1.554634</td>\n",
        "      <td> 0.245696</td>\n",
        "      <td> 0.599099</td>\n",
        "      <td> 0.778154</td>\n",
        "      <td> 0.718825</td>\n",
        "      <td> 0.443962</td>\n",
        "      <td> 1</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "<p>5 rows \u00d7 65 columns</p>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 11,
       "text": [
        "              Amplitude  AndersonDarling  Autocor_length  Beyond1Std  \\\n",
        "80.6708.5218    0.12700                1              44    0.278136   \n",
        "48.2742.1409    0.16525                1               1    0.185185   \n",
        "53.3970.353     0.14600                1              19    0.291367   \n",
        "14.8863.660     0.13200                1             146    0.316737   \n",
        "11.8751.968     0.48125                1              41    0.251832   \n",
        "\n",
        "              CAR_sigma     CAR_tau  CAR_tmean     Color       Con  \\\n",
        "80.6708.5218  -0.013622   37.158804  -0.195121 -0.356996  0.004307   \n",
        "48.2742.1409  -0.053792    3.563116  -1.814895 -0.520699  0.000000   \n",
        "53.3970.353   -0.003886  490.231454  -0.010910 -0.435427  0.000000   \n",
        "14.8863.660   -0.020506   23.280406  -0.291598 -0.260269  0.001062   \n",
        "11.8751.968   -0.016820  216.659786  -0.022690  0.386472  0.003670   \n",
        "\n",
        "                Eta_color     ...            Rcs      Skew  SlottedA_length  \\\n",
        "80.6708.5218  4478.196740     ...       0.275917  0.512233         0.026157   \n",
        "48.2742.1409    16.673054     ...       0.207167 -1.929378         1.024734   \n",
        "53.3970.353   2628.279811     ...       0.300025  0.387355        15.233322   \n",
        "14.8863.660   2390.816592     ...       0.372110 -0.547138       306.215875   \n",
        "11.8751.968    577.076785     ...       0.283297  0.464646         0.319352   \n",
        "\n",
        "              SmallKurtosis       Std  StetsonJ  StetsonK  StetsonK_AC  \\\n",
        "80.6708.5218       1.389718  0.063573  0.439091  0.763764     0.816338   \n",
        "48.2742.1409       7.272569  0.076998  0.006110  0.673988     0.721044   \n",
        "53.3970.353        0.318671  0.072600  0.399034  0.769279     0.765331   \n",
        "14.8863.660       -0.031463  0.072473  0.721305  0.814942     0.743765   \n",
        "11.8751.968        1.554634  0.245696  0.599099  0.778154     0.718825   \n",
        "\n",
        "              StetsonL  Class  \n",
        "80.6708.5218  0.274887      1  \n",
        "48.2742.1409  0.003726      1  \n",
        "53.3970.353   0.280160      1  \n",
        "14.8863.660   0.606026      1  \n",
        "11.8751.968   0.443962      1  \n",
        "\n",
        "[5 rows x 65 columns]"
       ]
      }
     ],
     "prompt_number": 11
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}