Enron POI Detector Project Assignment.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:83bab6926e5dbe8ceb741c048d732629e0b4702b777774839963f20fd616a7a7"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "heading",
     "level": 6,
     "metadata": {},
     "source": [
      "by Daria Alekseeva\n"
     ]
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Project Overview\n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "In 2000, Enron was one of the largest companies in the United States. By 2002, it had collapsed into bankruptcy due to widespread corporate fraud. In the resulting Federal investigation, there was a significant amount of typically confidential information entered into public record, including tens of thousands of emails and detailed financial data for top executives. \n",
      "\n",
      "In this project I will build a person of interest identifier based on financial and email data made public as a result of the Enron scandal. I use email and financial data for 146 executives at Enron to identify persons of interest in the fraud case.  A person of interest (POI) is someone who was indicted for fraud, settled with the government, or testified in exchange for immunity.  This report documents the machine learning techniques used in building a POI identifier.\n",
      "\n",
      "There are four major steps in my project:  \n",
      "1. Enron dataset\n",
      "2. Feature processing\n",
      "3. Algorithm\n",
      "4. Validation\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import os\n",
      "os.chdir('/home/daria/PycharmProjects/ud421-projects/final_project/')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from IPython.display import Image\n",
      "import matplotlib.pyplot as plt\n",
      "import sys\n",
      "import pickle\n",
      "from sklearn import preprocessing\n",
      "from time import time\n",
      "from sklearn.naive_bayes import GaussianNB\n",
      "from sklearn.metrics import accuracy_score\n",
      "from sklearn.grid_search import GridSearchCV\n",
      "from sklearn.metrics import precision_score\n",
      "from sklearn.metrics import recall_score\n",
      "sys.path.append(\"../tools/\")\n",
      "\n",
      "\n",
      "from feature_format import featureFormat\n",
      "from feature_format import targetFeatureSplit\n",
      "\n",
      "### features_list is a list of strings, each of which is a feature name\n",
      "### first feature must be \"poi\", as this will be singled out as the label\n",
      "features_list = [\"poi\"]\n",
      "\n",
      "### load the dictionary containing the dataset\n",
      "data_dict = pickle.load(open(\"final_project_dataset.pkl\", \"r\") )"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n"
       ]
      }
     ],
     "prompt_number": 14
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Firstly let's look how data is structured."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print \"There are \", len(data_dict.keys()), \" executives in Enron Dataset.\""
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "There are  146  executives in Enron Dataset.\n"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print data_dict.keys()\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "['METTS MARK', 'BAXTER JOHN C', 'ELLIOTT STEVEN', 'CORDES WILLIAM R', 'HANNON KEVIN P', 'MORDAUNT KRISTINA M', 'MEYER ROCKFORD G', 'MCMAHON JEFFREY', 'HORTON STANLEY C', 'PIPER GREGORY F', 'HUMPHREY GENE E', 'UMANOFF ADAM S', 'BLACHMAN JEREMY M', 'SUNDE MARTIN', 'GIBBS DANA R', 'LOWRY CHARLES P', 'COLWELL WESLEY', 'MULLER MARK S', 'JACKSON CHARLENE R', 'WESTFAHL RICHARD K', 'WALTERS GARETH W', 'WALLS JR ROBERT H', 'KITCHEN LOUISE', 'CHAN RONNIE', 'BELFER ROBERT', 'SHANKMAN JEFFREY A', 'WODRASKA JOHN', 'BERGSIEKER RICHARD P', 'URQUHART JOHN A', 'BIBI PHILIPPE A', 'RIEKER PAULA H', 'WHALEY DAVID A', 'BECK SALLY W', 'HAUG DAVID L', 'ECHOLS JOHN B', 'MENDELSOHN JOHN', 'HICKERSON GARY J', 'CLINE KENNETH W', 'LEWIS RICHARD', 'HAYES ROBERT E', 'MCCARTY DANNY J', 'KOPPER MICHAEL J', 'LEFF DANIEL P', 'LAVORATO JOHN J', 'BERBERIAN DAVID', 'DETMERING TIMOTHY J', 'WAKEHAM JOHN', 'POWERS WILLIAM', 'GOLD JOSEPH', 'BANNANTINE JAMES M', 'DUNCAN JOHN H', 'SHAPIRO RICHARD S', 'SHERRIFF JOHN R', 'SHELBY REX', 'LEMAISTRE CHARLES', 'DEFFNER JOSEPH M', 'KISHKILL JOSEPH G', 'WHALLEY LAWRENCE G', 'MCCONNELL MICHAEL S', 'PIRO JIM', 'DELAINEY DAVID W', 'SULLIVAN-SHAKLOVITZ COLLEEN', 'WROBEL BRUCE', 'LINDHOLM TOD A', 'MEYER JEROME J', 'LAY KENNETH L', 'BUTTS ROBERT H', 'OLSON CINDY K', 'MCDONALD REBECCA', 'CUMBERLAND MICHAEL S', 'GAHN ROBERT S', 'MCCLELLAN GEORGE', 'HERMANN ROBERT J', 'SCRIMSHAW MATTHEW', 'GATHMANN WILLIAM D', 'HAEDICKE MARK E', 'BOWEN JR RAYMOND M', 'GILLIS JOHN', 'FITZGERALD JAY L', 'MORAN MICHAEL P', 'REDMOND BRIAN L', 'BAZELIDES PHILIP J', 'BELDEN TIMOTHY N', 'DURAN WILLIAM D', 'THORN TERENCE H', 'FASTOW ANDREW S', 'FOY JOE', 'CALGER CHRISTOPHER F', 'RICE KENNETH D', 'KAMINSKI WINCENTY J', 'LOCKHART EUGENE E', 'COX DAVID', 'OVERDYKE JR JERE C', 'PEREIRA PAULO V. FERRAZ', 'STABLER FRANK', 'SKILLING JEFFREY K', 'BLAKE JR. NORMAN P', 'SHERRICK JEFFREY B', 'PRENTICE JAMES', 'GRAY RODNEY', 'PICKERING MARK R', 'THE TRAVEL AGENCY IN THE PARK', 'NOLES JAMES L', 'KEAN STEVEN J', 'TOTAL', 'FOWLER PEGGY', 'WASAFF GEORGE', 'WHITE JR THOMAS E', 'CHRISTODOULOU DIOMEDES', 'ALLEN PHILLIP K', 'SHARP VICTORIA T', 'JAEDICKE ROBERT', 'WINOKUR JR. HERBERT S', 'BROWN MICHAEL', 'BADUM JAMES P', 'HUGHES JAMES A', 'REYNOLDS LAWRENCE', 'DIMICHELE RICHARD G', 'BHATNAGAR SANJAY', 'CARTER REBECCA C', 'BUCHANAN HAROLD G', 'YEAP SOON', 'MURRAY JULIA H', 'GARLAND C KEVIN', 'DODSON KEITH', 'YEAGER F SCOTT', 'HIRKO JOSEPH', 'DIETRICH JANET R', 'DERRICK JR. JAMES V', 'FREVERT MARK A', 'PAI LOU L', 'BAY FRANKLIN R', 'HAYSLETT RODERICK J', 'FUGH JOHN L', 'FALLON JAMES B', 'KOENIG MARK E', 'SAVAGE FRANK', 'IZZO LAWRENCE L', 'TILNEY ELIZABETH A', 'MARTIN AMANDA K', 'BUY RICHARD B', 'GRAMM WENDY L', 'CAUSEY RICHARD A', 'TAYLOR MITCHELL S', 'DONAHUE JR JEFFREY M', 'GLISAN JR BEN F']\n"
       ]
      }
     ],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print data_dict['BUY RICHARD B']\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "{'salary': 330546, 'to_messages': 3523, 'deferral_payments': 649584, 'total_payments': 2355702, 'exercised_stock_options': 2542813, 'bonus': 900000, 'restricted_stock': 901657, 'shared_receipt_with_poi': 2333, 'restricted_stock_deferred': 'NaN', 'total_stock_value': 3444470, 'expenses': 'NaN', 'loan_advances': 'NaN', 'from_messages': 1053, 'other': 400572, 'from_this_person_to_poi': 71, 'poi': False, 'director_fees': 'NaN', 'deferred_income': -694862, 'long_term_incentive': 769862, 'email_address': 'rick.buy@enron.com', 'from_poi_to_this_person': 156}\n"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "The Enron Data"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "First of all I\u2019d like to have a look at my data and check it for outliers. I plot salaries and bonuses on Enron employees and see an outlier in the data."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "features = [\"salary\", \"bonus\"]\n",
      "#data_dict.pop('TOTAL', 0)\n",
      "data = featureFormat(data_dict, features)\n",
      "### plot features\n",
      "for point in data:\n",
      "    salary = point[0]\n",
      "    bonus = point[1]\n",
      "    plt.scatter( salary, bonus )\n",
      "\n",
      "plt.xlabel(\"salary\")\n",
      "plt.ylabel(\"bonus\")\n",
      "#plt.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 6,
       "text": [
        "<matplotlib.text.Text at 0x7f882864d250>"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "Image(filename='data.png')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAisAAAGtCAYAAADXrzRwAAAgAElEQVR4Xu2dB5gVVdKGa4YgipgQ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\nCCBWVkeI9yEAAQhAAAIQSJXA/wPUopVf/RVQEwAAAABJRU5ErkJggg==\n",
       "prompt_number": 7,
       "text": [
        "<IPython.core.display.Image at 0x7f88502fd610>"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "When I check it I see this is a number for total salary and bonus. As this is not sensible information for our analysis I remove it manually. Two more outliers (SKILLING JEFFREY and LAY KENNETH) I keep in dataset as these values real and actually they are already a sign of these two managers being involved in the fraud. Now dataset look like this:\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "### remove any outliers before proceeding further\n",
      "features = [\"salary\", \"bonus\"]\n",
      "data_dict.pop('TOTAL', 0)\n",
      "data = featureFormat(data_dict, features)\n",
      "\n",
      "### remove NAN's from dataset\n",
      "outliers = []\n",
      "for key in data_dict:\n",
      "    val = data_dict[key]['salary']\n",
      "    if val == 'NaN':\n",
      "        continue\n",
      "    outliers.append((key, int(val)))\n",
      "\n",
      "outliers_final = (sorted(outliers,key=lambda x:x[1],reverse=True)[:4])\n",
      "### print top 4 salaries\n",
      "print outliers_final"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[('SKILLING JEFFREY K', 1111258), ('LAY KENNETH L', 1072321), ('FREVERT MARK A', 1060932), ('PICKERING MARK R', 655037)]\n"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "Image(filename='data_no_outliers.png')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "png": "iVBORw0KGgoAAAANSUhEUgAAApEAAAHpCAYAAADasmf6AAAgAElEQVR4XuydB3hUxdfG302nhSao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\ncgUBBj3jAAAAAElFTkSuQmCC\n",
       "prompt_number": 9,
       "text": [
        "<IPython.core.display.Image at 0x7f88502fd790>"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Feature processing\n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "After cleaning the data from outliers I had to pick the most sensible features to use. First I picked 'from_poi_to_this_person' and  'from_this_person_to_poi' but there is was no strong pattern when I plotted the data so I used fractions for both features of \u201cfrom/to poi messages\u201d and \u201ctotal from/to messages\u201d. \n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "### create new features\n",
      "### new features are: fraction_to_poi_email,fraction_from_poi_email\n",
      "\n",
      "def dict_to_list(key,normalizer):\n",
      "    new_list=[]\n",
      "\n",
      "    for i in data_dict:\n",
      "        if data_dict[i][key]==\"NaN\" or data_dict[i][normalizer]==\"NaN\":\n",
      "            new_list.append(0.)\n",
      "        elif data_dict[i][key]>=0:\n",
      "            new_list.append(float(data_dict[i][key])/float(data_dict[i][normalizer]))\n",
      "    return new_list\n",
      "\n",
      "### create two lists of new features\n",
      "fraction_from_poi_email=dict_to_list(\"from_poi_to_this_person\",\"to_messages\")\n",
      "fraction_to_poi_email=dict_to_list(\"from_this_person_to_poi\",\"from_messages\")\n",
      "\n",
      "### insert new features into data_dict\n",
      "count=0\n",
      "for i in data_dict:\n",
      "    data_dict[i][\"fraction_from_poi_email\"]=fraction_from_poi_email[count]\n",
      "    data_dict[i][\"fraction_to_poi_email\"]=fraction_to_poi_email[count]\n",
      "    count +=1\n",
      "\n",
      "    \n",
      "features_list = [\"poi\", \"fraction_from_poi_email\", \"fraction_to_poi_email\"]    \n",
      "    ### store to my_dataset for easy export below\n",
      "my_dataset = data_dict\n",
      "\n",
      "\n",
      "### these two lines extract the features specified in features_list\n",
      "### and extract them from data_dict, returning a numpy array\n",
      "data = featureFormat(my_dataset, features_list)\n",
      "\n",
      "### plot new features\n",
      "for point in data:\n",
      "    from_poi = point[1]\n",
      "    to_poi = point[2]\n",
      "    plt.scatter( from_poi, to_poi )\n",
      "    if point[0] == 1:\n",
      "        plt.scatter(from_poi, to_poi, color=\"r\", marker=\"*\")\n",
      "plt.xlabel(\"fraction of emails this person gets from poi\")\n",
      "plt.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 24
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "Image(filename='data_poi_nonpoi.png')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAfcAAAGBCAYAAAB7Dvj7AAAgAElEQVR4XuydB3gU1RbHfzNbsukh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\nAAAASUVORK5CYII=\n",
       "prompt_number": 47,
       "text": [
        "<IPython.core.display.Image at 0x7fa87e2c71d0>"
       ]
      }
     ],
     "prompt_number": 47
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "In order to find the most effective features for classification, feature selection using \u201cDecision Tree\u201d  was deployed to rank the features. Selection features was half manual iterative process. First I put all the possible features into features_list and then started deleting them one by one using score value and human intuition.\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "features_list = [\"poi\", \"salary\", \"bonus\", \"fraction_from_poi_email\", \"fraction_to_poi_email\",\n",
      "                 'deferral_payments', 'total_payments', 'loan_advances', 'restricted_stock_deferred',\n",
      "                 'deferred_income', 'total_stock_value', 'expenses', 'exercised_stock_options',\n",
      "                 'long_term_incentive', 'shared_receipt_with_poi', 'restricted_stock', 'director_fees']\n",
      "data = featureFormat(my_dataset, features_list)\n",
      "\n",
      "### split into labels and features (this line assumes that the first\n",
      "### feature in the array is the label, which is why \"poi\" must always\n",
      "### be first in features_list\n",
      "labels, features = targetFeatureSplit(data)\n",
      "\n",
      "### split data into training and testing datasets\n",
      "from sklearn import cross_validation\n",
      "features_train, features_test, labels_train, labels_test = cross_validation.train_test_split(features, labels, test_size=0.1, random_state=42)\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "from sklearn.tree import DecisionTreeClassifier\n",
      "\n",
      "t0 = time()\n",
      "\n",
      "clf = DecisionTreeClassifier()\n",
      "clf.fit(features_train,labels_train)\n",
      "score = clf.score(features_test,labels_test)\n",
      "pred= clf.predict(features_test)\n",
      "print 'accuracy', score\n",
      "\n",
      "print \"Decision tree algorithm time:\", round(time()-t0, 3), \"s\"\n",
      "\n",
      "\n",
      "\n",
      "importances = clf.feature_importances_\n",
      "import numpy as np\n",
      "indices = np.argsort(importances)[::-1]\n",
      "print 'Feature Ranking: '\n",
      "for i in range(16):\n",
      "    print \"{} feature {} ({})\".format(i+1,features_list[i+1],importances[indices[i]])\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "accuracy 0.8\n",
        "Decision tree algorithm time: 0.003 s\n",
        "Feature Ranking: \n",
        "1 feature salary (0.211707133446)\n",
        "2 feature bonus (0.181365499334)\n",
        "3 feature fraction_from_poi_email (0.14622972935)\n",
        "4 feature fraction_to_poi_email (0.120901730257)\n",
        "5 feature deferral_payments (0.118337314859)\n",
        "6 feature total_payments (0.0879795396419)\n",
        "7 feature loan_advances (0.0747826086957)\n",
        "8 feature restricted_stock_deferred (0.0377115287109)\n",
        "9 feature deferred_income (0.0209849157054)\n",
        "10 feature total_stock_value (0.0)\n",
        "11 feature expenses (0.0)\n",
        "12 feature exercised_stock_options (0.0)\n",
        "13 feature long_term_incentive (0.0)\n",
        "14 feature shared_receipt_with_poi (0.0)\n",
        "15 feature restricted_stock (0.0)\n",
        "16 feature director_fees (0.0)\n"
       ]
      }
     ],
     "prompt_number": 20
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Finally I picked 10 features which are:\n",
      "```\n",
      "[\"salary\", \"bonus\", \"fraction_from_poi_email\", \"fraction_to_poi_email\", 'deferral_payments', 'total_payments', 'loan_advances', 'restricted_stock_deferred', 'deferred_income', 'total_stock_value']\n",
      "```\n",
      "\n",
      "Accuracy for this feature set is around **0.8**. \n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "But with these features my precision and recall were too low (less than 0.3) so I had to change my strategy and manually pick features which gave me precision and recall values over 0.3. In this dataset I cannot use accuracy for evaluating my algorithm because there a few POI\u2019s in dataset and the best evaluator are precision and recall. There were only 18 examples of POIs in the dataset.  There were 35 people who were POIs in \u201creal life\u201d, but for various reasons, half of those are not present in this dataset.  \n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Finally I picked the following features:\n",
      "**[\"fraction_from_poi_email\", \"fraction_to_poi_email\", \u201cshared_receipt_with_poi\u201d]**\n"
     ]
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Algorithm Selection and Tuning\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "features_list = [\"poi\", \"fraction_from_poi_email\", \"fraction_to_poi_email\", \"shared_receipt_with_poi\"]\n",
      "\n",
      "### try Naive Bayes for prediction\n",
      "t0 = time()\n",
      "\n",
      "clf = GaussianNB()\n",
      "clf.fit(features_train, labels_train)\n",
      "pred = clf.predict(features_test)\n",
      "accuracy = accuracy_score(pred,labels_test)\n",
      "print accuracy\n",
      "\n",
      "print \"NB algorithm time:\", round(time()-t0, 3), \"s\""
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "0.2\n",
        "NB algorithm time: 0.053 s\n"
       ]
      }
     ],
     "prompt_number": 18
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Firstly I tried Naive Bayes accuracy was lower than with Decision Tree Algorithm (0.83 and 0.9 respectively). I made a conclusion that that the feature set I used does not suit the distributional and interactive assumptions of Naive Bayes well enough. \n",
      "I selected Decision Tree Algorithm for the POI identifier. It gave me accuracy before tuning parameters = 0.9. No feature scaling was deployed, as it\u2019s not necessary when using a decision tree.\n",
      "After selecting features and algorithm I manually tuned parameter min_samples_split. \n",
      "\n",
      "```\n",
      "min_samples_split    precision    recall\n",
      "      2                0.67        0.8\n",
      "      3                0.57        0.8\n",
      "      4                0.57        0.8\n",
      "      5                0.8         0.8\n",
      "      6                0.8         0.8\n",
      "      7                0.67        0.8\n",
      "   average             0.68        0.8\n",
      "```   \n",
      "\n",
      "It turned out that the best values for min_samples_split are 5 and 6.\n"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "### use manual tuning parameter min_samples_split\n",
      "clf = DecisionTreeClassifier(min_samples_split=5)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 22
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Analysis Validation and Performance\n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "This process was validated using 3-fold cross-validation, precision and recall scores.\n",
      "First I used accuracy to evaluate my algorithm. It was a mistake because in this case we have a class imbalance problem - the number of POIs is small compared to the total number of examples in the dataset. So I had to use precision and recall for these activities instead. \n",
      "I was able to reach average value of precision = 0.68, recall =  0.8."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "### features_list is a list of strings, each of which is a feature name\n",
      "### first feature must be \"poi\", as this will be singled out as the label\n",
      "features_list = [\"poi\", \"fraction_from_poi_email\", \"fraction_to_poi_email\", 'shared_receipt_with_poi']\n",
      "\n",
      "\n",
      "### store to my_dataset for easy export below\n",
      "my_dataset = data_dict\n",
      "\n",
      "\n",
      "### these two lines extract the features specified in features_list\n",
      "### and extract them from data_dict, returning a numpy array\n",
      "data = featureFormat(my_dataset, features_list)\n",
      "\n",
      "\n",
      "### split into labels and features (this line assumes that the first\n",
      "### feature in the array is the label, which is why \"poi\" must always\n",
      "### be first in features_list\n",
      "labels, features = targetFeatureSplit(data)\n",
      "\n",
      "\n",
      "### machine learning goes here!\n",
      "### please name your classifier clf for easy export below\n",
      "\n",
      "### deploying feature selection\n",
      "from sklearn import cross_validation\n",
      "features_train, features_test, labels_train, labels_test = cross_validation.train_test_split(features, labels, test_size=0.1, random_state=42)\n",
      "\n",
      "### use KFold for split and validate algorithm\n",
      "from sklearn.cross_validation import KFold\n",
      "kf=KFold(len(labels),3)\n",
      "for train_indices, test_indices in kf:\n",
      "    #make training and testing sets\n",
      "    features_train= [features[ii] for ii in train_indices]\n",
      "    features_test= [features[ii] for ii in test_indices]\n",
      "    labels_train=[labels[ii] for ii in train_indices]\n",
      "    labels_test=[labels[ii] for ii in test_indices]\n",
      "\n",
      "from sklearn.tree import DecisionTreeClassifier\n",
      "\n",
      "t0 = time()\n",
      "\n",
      "clf = DecisionTreeClassifier()\n",
      "clf.fit(features_train,labels_train)\n",
      "score = clf.score(features_test,labels_test)\n",
      "print 'accuracy before tuning ', score\n",
      "\n",
      "print \"Decision tree algorithm time:\", round(time()-t0, 3), \"s\"\n",
      "\n",
      "\n",
      "### use manual tuning parameter min_samples_split\n",
      "t0 = time()\n",
      "clf = DecisionTreeClassifier(min_samples_split=5)\n",
      "clf = clf.fit(features_train,labels_train)\n",
      "pred= clf.predict(features_test)\n",
      "print(\"done in %0.3fs\" % (time() - t0))\n",
      "\n",
      "acc=accuracy_score(labels_test, pred)\n",
      "\n",
      "print \"Validating algorithm:\"\n",
      "print \"accuracy after tuning = \", acc\n",
      "\n",
      "# function for calculation ratio of true positives\n",
      "# out of all positives (true + false)\n",
      "print 'precision = ', precision_score(labels_test,pred)\n",
      "\n",
      "# function for calculation ratio of true positives\n",
      "# out of true positives and false negatives\n",
      "print 'recall = ', recall_score(labels_test,pred)\n",
      "\n",
      "\n",
      "### dump your classifier, dataset and features_list so\n",
      "### anyone can run/check your results\n",
      "pickle.dump(clf, open(\"my_classifier.pkl\", \"w\") )\n",
      "pickle.dump(data_dict, open(\"my_dataset.pkl\", \"w\") )\n",
      "pickle.dump(features_list, open(\"my_feature_list.pkl\", \"w\") )"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "accuracy before tuning  0.866666666667\n",
        "Decision tree algorithm time: 0.001 s\n",
        "done in 0.001s\n",
        "Validating algorithm:\n",
        "accuracy after tuning =  0.933333333333\n",
        "precision =  0.8\n",
        "recall =  0.8\n"
       ]
      }
     ],
     "prompt_number": 26
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Discussion and Conclusions\n"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "The precision can be interpreted as the likelihood that a person who is identified as a POI is actually a true POI; the fact that this is 0.68 means that using this identifier to flag POI\u2019s would result in 32% of the positive flags being false alarms. Recall measures how likely it is that identifier will flag a POI in the test set. 80% of the time it would catch that person, and 20% of the time it wouldn\u2019t.  \n",
      "These numbers are quite good but we still can improve the strategy. One of the possible paths to improvement is digging in to the emails data more. The email features in the starter dataset were aggregated over all the messages for a given person. By digging into the text of each individual\u2019s messages, it\u2019s possible that more detailed patterns (say, messages to/from a specific address, rather than just messages to/from any POI address, or the usage of specific vocabulary terms) might emerge.  Since we live in a world in which more POI finance data might not be easy to find, the next realistic thing to try might be to extract more data from the emails. \n"
     ]
    }
   ],
   "metadata": {}
  }
 ]
}