5. Scikit Learn - Distance Functions and Scoring Functions.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:1114cddd4de8e3cdbc294e4080f9c002f75f7be41ab26055a1062d16decf797a"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## Distance and Scoring Functions\n",
      "If people asked me to name one thing that has been constantly ignored or not given importance in machine learning, first thing\n",
      "from top of my head would be evaluation. That is partially because some contexts(search, recommender systems) it is somehow difficult to build an evaluation function that first perfectly aligns with the business logic. Partially, people end up measuring and optimizing towards wrong thing or not the _best_ thing for the business problem. \n",
      "\n",
      "Consider the following scenario; you are predicting the number of items sold for a variety of items in a distribution problem. Assume that most of the businesses that require these items are small businesses, so they require small number of items. Large business require large number of items but comparing to small businesses their number is much less. Further, assume that errors in the number of items sold in the prediction would result in a big problem for small businesses as they struggle to sell. However, large number of items in error for large businesses may not cause a big problem for them as they could sell somehow items easier than small businesses.\n",
      "\n",
      "For this application $l_2$(RMSE) may not make sense, because you will have errors in small quantitites and even if you are optimizing total error to be smallest, you are not optimizing for the business logic. $l_2$ results in small errors in large number of observations(small errors for both large businesses and small businesses).\n",
      "\n",
      "$l_1$ could be better evaluation distance function when you compute the error, even if its value may exceed $l_2$. $l_1$ produces rather sparse error distribution and if you penalize the number of items to error, it will shrink the error to large number of businesses which is an optimal state for the business logic. \n",
      "\n",
      "At this point, one needs to also think, maybe the evaluation function should not be the number of items to be sold, but rather $\\sum p(items * \\text{probability of an item is sold } |\\text{ business}_i)$. \n",
      "\n",
      "Of course, evaluation does not really mean anything without a proper evaluation function. If you are predicting $\\hat{y}$ and want to measure cost function from $y$, distance function that you will use is crucially important as it will determine the __model__. So, the distance function as well as what type of scoring function is using in the evaluation is quite important for how your model performs."
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Luckily, Scikit-Learn comes with a lot of distance and scoring functions out of the box. If you want another distance and scoring function, separate from these, it is easy to build one based on your needs as well. This makes it very convenient to try different scoring methods when you do the model selection. "
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Evaluation Methods for Unsupervised Learning Algorithm\n",
      "It is relatively harder problem to evaluate a learning algorithm in an unsupervised setting as there is no ground truth labels for each class and model selection is not only concerned by the model itself but also the number of classes that the model is setting in the learning algorithm. \n",
      "\n",
      "As in the supervised learning algorithm case, Scikit-Learn provides a variety of scoring functions out of the box to compute various metrics for a given clustering algorithm or compare multiple ones based on a single metric and preferably same number of clusters. `homogeneity_score`, `completeness_score` and `v_measure_score`(harmonic mean of `homogeneity_score` and `completeness_score`) are examples of these scores."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "% matplotlib inline\n",
      "import itertools\n",
      "import matplotlib.pyplot as plt\n",
      "import os\n",
      "import pandas as pd\n",
      "import seaborn as sns\n",
      "from sklearn import cluster\n",
      "from sklearn import cross_validation\n",
      "from sklearn import datasets\n",
      "from sklearn import ensemble\n",
      "from sklearn import linear_model\n",
      "from sklearn import metrics\n",
      "from sklearn import naive_bayes\n",
      "from sklearn import preprocessing\n",
      "from sklearn import svm\n",
      "from sklearn import tree\n",
      "\n",
      "plt.style.use('fivethirtyeight')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Scoring Functions"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "#### Custom Metric Functions"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "metrics.SCORERS"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 2,
       "text": [
        "{'accuracy': make_scorer(accuracy_score),\n",
        " 'adjusted_rand_score': make_scorer(adjusted_rand_score),\n",
        " 'average_precision': make_scorer(average_precision_score, needs_threshold=True),\n",
        " 'f1': make_scorer(f1_score),\n",
        " 'f1_macro': <repr(<sklearn.metrics.scorer._PredictScorer at 0x10970c110>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'f1_micro': <repr(<sklearn.metrics.scorer._PredictScorer at 0x10970c150>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'f1_samples': <repr(<sklearn.metrics.scorer._PredictScorer at 0x10970c190>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'f1_weighted': <repr(<sklearn.metrics.scorer._PredictScorer at 0x10970c1d0>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'log_loss': make_scorer(log_loss, greater_is_better=False, needs_proba=True),\n",
        " 'mean_absolute_error': make_scorer(mean_absolute_error, greater_is_better=False),\n",
        " 'mean_squared_error': make_scorer(mean_squared_error, greater_is_better=False),\n",
        " 'median_absolute_error': make_scorer(median_absolute_error, greater_is_better=False),\n",
        " 'precision': make_scorer(precision_score),\n",
        " 'precision_macro': <repr(<sklearn.metrics.scorer._PredictScorer at 0x109700e50>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'precision_micro': <repr(<sklearn.metrics.scorer._PredictScorer at 0x109700e90>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'precision_samples': <repr(<sklearn.metrics.scorer._PredictScorer at 0x109700ed0>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'precision_weighted': <repr(<sklearn.metrics.scorer._PredictScorer at 0x109700f10>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'r2': make_scorer(r2_score),\n",
        " 'recall': make_scorer(recall_score),\n",
        " 'recall_macro': <repr(<sklearn.metrics.scorer._PredictScorer at 0x109700f90>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'recall_micro': <repr(<sklearn.metrics.scorer._PredictScorer at 0x109700fd0>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'recall_samples': <repr(<sklearn.metrics.scorer._PredictScorer at 0x10970c050>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'recall_weighted': <repr(<sklearn.metrics.scorer._PredictScorer at 0x10970c090>) failed: AttributeError: 'functools.partial' object has no attribute '__name__'>,\n",
        " 'roc_auc': make_scorer(roc_auc_score, needs_threshold=True)}"
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "scoring_funcs = [ii for ii in dir(metrics) if 'score' in ii and not ii.startswith('_')]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "scoring_funcs"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 4,
       "text": [
        "['accuracy_score',\n",
        " 'adjusted_mutual_info_score',\n",
        " 'adjusted_rand_score',\n",
        " 'average_precision_score',\n",
        " 'brier_score_loss',\n",
        " 'completeness_score',\n",
        " 'consensus_score',\n",
        " 'explained_variance_score',\n",
        " 'f1_score',\n",
        " 'fbeta_score',\n",
        " 'get_scorer',\n",
        " 'homogeneity_score',\n",
        " 'jaccard_similarity_score',\n",
        " 'label_ranking_average_precision_score',\n",
        " 'make_scorer',\n",
        " 'mutual_info_score',\n",
        " 'normalized_mutual_info_score',\n",
        " 'precision_recall_fscore_support',\n",
        " 'precision_score',\n",
        " 'r2_score',\n",
        " 'recall_score',\n",
        " 'roc_auc_score',\n",
        " 'scorer',\n",
        " 'silhouette_score',\n",
        " 'v_measure_score']"
       ]
      }
     ],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "_DATA_DIR = 'data'\n",
      "_TRAINING_DATA_PATH = os.path.join(_DATA_DIR, 'userart-mat-training.csv')\n",
      "_TEST_DATA_PATH = os.path.join(_DATA_DIR, 'userart-mat-test.csv')\n",
      "_ARTIST_TAGS_PATH = os.path.join(_DATA_DIR, 'artists-tags.txt')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 5
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "df = pd.read_csv(_TRAINING_DATA_PATH)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "df.head(10)"
     ],
     "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>ArtistID</th>\n",
        "      <th>user_000001</th>\n",
        "      <th>user_000002</th>\n",
        "      <th>user_000003</th>\n",
        "      <th>user_000004</th>\n",
        "      <th>user_000005</th>\n",
        "      <th>user_000006</th>\n",
        "      <th>user_000007</th>\n",
        "      <th>user_000008</th>\n",
        "      <th>user_000009</th>\n",
        "      <th>...</th>\n",
        "      <th>user_000890</th>\n",
        "      <th>user_000891</th>\n",
        "      <th>user_000892</th>\n",
        "      <th>user_000893</th>\n",
        "      <th>user_000894</th>\n",
        "      <th>user_000895</th>\n",
        "      <th>user_000896</th>\n",
        "      <th>user_000897</th>\n",
        "      <th>user_000898</th>\n",
        "      <th>user_000899</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>0</th>\n",
        "      <td> c489cd1c-d8d3-4dcc-b7b9-dcf7cd95d2ed</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1</th>\n",
        "      <td> 9ccbb935-33fd-4df8-ae2d-779497c5630a</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>2</th>\n",
        "      <td> c71b2cc1-2737-4f76-bbf2-9ee505b4d90d</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>3</th>\n",
        "      <td> 0aa8294b-6332-4b65-b677-e3a1f8591d3b</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>4</th>\n",
        "      <td> 67673557-2310-41d3-83b9-e6e0cb1d65d5</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>5</th>\n",
        "      <td> 69ead53e-b2d4-45f4-b777-71363fdeff28</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>6</th>\n",
        "      <td> 5a092e7c-faa9-45d0-88c6-052205804e6c</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>7</th>\n",
        "      <td> 81fdb05d-3f73-4c38-bf69-d3c642a04439</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>8</th>\n",
        "      <td> b0f475fe-ad28-4cf8-92f0-b5ca978004b1</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 1</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>9</th>\n",
        "      <td> 6994f98a-a41e-4d74-af5b-9e3fab3c19b4</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td>...</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "      <td> 0</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "<p>10 rows \u00d7 847 columns</p>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 7,
       "text": [
        "                               ArtistID  user_000001  user_000002  \\\n",
        "0  c489cd1c-d8d3-4dcc-b7b9-dcf7cd95d2ed            0            0   \n",
        "1  9ccbb935-33fd-4df8-ae2d-779497c5630a            0            0   \n",
        "2  c71b2cc1-2737-4f76-bbf2-9ee505b4d90d            0            0   \n",
        "3  0aa8294b-6332-4b65-b677-e3a1f8591d3b            0            0   \n",
        "4  67673557-2310-41d3-83b9-e6e0cb1d65d5            0            0   \n",
        "5  69ead53e-b2d4-45f4-b777-71363fdeff28            0            0   \n",
        "6  5a092e7c-faa9-45d0-88c6-052205804e6c            0            0   \n",
        "7  81fdb05d-3f73-4c38-bf69-d3c642a04439            0            0   \n",
        "8  b0f475fe-ad28-4cf8-92f0-b5ca978004b1            0            0   \n",
        "9  6994f98a-a41e-4d74-af5b-9e3fab3c19b4            0            0   \n",
        "\n",
        "   user_000003  user_000004  user_000005  user_000006  user_000007  \\\n",
        "0            0            0            0            0            0   \n",
        "1            0            0            0            0            0   \n",
        "2            0            0            0            0            0   \n",
        "3            0            0            0            0            0   \n",
        "4            0            0            0            0            0   \n",
        "5            0            0            0            0            0   \n",
        "6            0            0            0            0            0   \n",
        "7            0            0            0            0            0   \n",
        "8            0            0            0            0            0   \n",
        "9            0            0            0            0            0   \n",
        "\n",
        "   user_000008  user_000009     ...       user_000890  user_000891  \\\n",
        "0            0            0     ...                 0            0   \n",
        "1            0            0     ...                 0            0   \n",
        "2            0            0     ...                 0            0   \n",
        "3            0            0     ...                 0            0   \n",
        "4            0            0     ...                 0            0   \n",
        "5            0            0     ...                 0            0   \n",
        "6            0            0     ...                 0            0   \n",
        "7            0            0     ...                 0            0   \n",
        "8            0            0     ...                 0            0   \n",
        "9            0            0     ...                 0            0   \n",
        "\n",
        "   user_000892  user_000893  user_000894  user_000895  user_000896  \\\n",
        "0            0            0            0            0            0   \n",
        "1            0            0            0            0            0   \n",
        "2            0            0            0            0            0   \n",
        "3            0            0            0            0            0   \n",
        "4            0            0            0            0            0   \n",
        "5            0            0            0            0            0   \n",
        "6            0            0            0            0            0   \n",
        "7            0            0            0            0            0   \n",
        "8            1            0            0            0            0   \n",
        "9            0            0            0            0            0   \n",
        "\n",
        "   user_000897  user_000898  user_000899  \n",
        "0            0            0            0  \n",
        "1            0            0            0  \n",
        "2            0            1            0  \n",
        "3            0            0            0  \n",
        "4            0            0            0  \n",
        "5            1            0            0  \n",
        "6            0            0            0  \n",
        "7            0            0            0  \n",
        "8            0            0            0  \n",
        "9            0            0            0  \n",
        "\n",
        "[10 rows x 847 columns]"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "del df['ArtistID']"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "X = df.as_matrix().astype(float)\n",
      "X.shape"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 9,
       "text": [
        "(17119, 846)"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "n_artists, n_users = X.shape\n",
      "frame_number = n_artists // n_users + 1"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 10
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "frame_number"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 11,
       "text": [
        "21"
       ]
      }
     ],
     "prompt_number": 11
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "col_number = row_number = 5\n",
      "fig, ax = plt.subplots(row_number, col_number, figsize=(16, 12))\n",
      "for ii in range(25):\n",
      "    row, col = ii // row_number, ii % row_number\n",
      "    indices = range((row-1)*n_users, row * n_users)\n",
      "    ax[row, col].imshow(X[indices], cmap='bone')\n",
      "    ax[row, col].set_xticks([])\n",
      "    ax[row, col].set_yticks([])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAA7IAAALbCAYAAADHDLzSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x104aaf650>"
       ]
      }
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Internal Evaluation"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "silhouette_scores = []\n",
      "n_clusters = range(2, 15)\n",
      "\n",
      "for n_cluster in n_clusters:\n",
      "    print('K-Means run for number of clusters: {} is started'.format(n_cluster))\n",
      "    kmeans = cluster.KMeans(n_clusters=n_cluster, init=\"k-means++\").fit(X)\n",
      "    labels = kmeans.labels_\n",
      "    silhouette_scores.append(metrics.silhouette_score(X, labels, metric='euclidean'))\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "K-Means run for number of clusters: 2 is started\n",
        "K-Means run for number of clusters: 3 is started"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "K-Means run for number of clusters: 4 is started"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "K-Means run for number of clusters: 5 is started"
       ]
      }
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plt.figure(figsize=(16, 12))\n",
      "plt.plot(n_clusters, silhouette_scores)\n",
      "plt.ylabel(\"Silhouette Score\");\n",
      "plt.xlabel(\"Cluster\");\n",
      "plt.title(\"Silhouette vs Cluster Size\");"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### External Evaluation"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "iris = datasets.load_iris()\n",
      "X = iris.data\n",
      "y = iris.target"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plt.figure(figsize=(16, 12))\n",
      "plt.imshow(X.reshape((30, 20)), cmap='bone')\n",
      "plt.xticks([]);\n",
      "plt.yticks([]);"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "n_clusters = range(2, 15)\n",
      "adjusted_random_scores = []\n",
      "v_measure_scores = []\n",
      "\n",
      "for n_cluster in n_clusters:\n",
      "    print('K-Means run for number of clusters: {} is started'.format(n_cluster))\n",
      "    kmeans = cluster.KMeans(n_clusters=n_cluster, init=\"k-means++\").fit(X)\n",
      "    labels = kmeans.labels_\n",
      "    adjusted_random_scores.append(metrics.adjusted_rand_score(y, labels) )\n",
      "    v_measure_scores.append(metrics.v_measure_score(y, labels))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plt.figure(figsize=(16, 12))\n",
      "plt.plot(n_clusters, adjusted_random_scores)\n",
      "plt.ylabel(\"Adjusted Random Score\");\n",
      "plt.xlabel(\"Cluster\");\n",
      "plt.title(\"Adjusted Random Score vs Cluster Size\");"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plt.figure(figsize=(16, 12))\n",
      "plt.plot(n_clusters, v_measure_scores)\n",
      "plt.ylabel(\"V Measure Score\");\n",
      "plt.xlabel(\"Cluster\");\n",
      "plt.title(\"V Measure Score vs Cluster Size\");"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Supervised Setting"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "df = sns.load_dataset(\"titanic\")\n",
      "df.head(10)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "df.dropna(inplace=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Turn all the columns to categorical variables!"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "label_encoder = preprocessing.LabelEncoder()\n",
      "df['embarked'] = label_encoder.fit_transform(df.embarked)\n",
      "df['embark_town'] = label_encoder.fit_transform(df.embark_town)\n",
      "df['alive'] = label_encoder.fit_transform(df.alive)\n",
      "df['alone'] = label_encoder.fit_transform(df.alone)\n",
      "df['deck'] = label_encoder.fit_transform(df.deck)\n",
      "df['adult_male'] = label_encoder.fit_transform(df.adult_male)\n",
      "df['who'] = label_encoder.fit_transform(df.who)\n",
      "df['class'] = label_encoder.fit_transform(df['class'])\n",
      "df['parch'] = label_encoder.fit_transform(df['parch'])\n",
      "df['sex'] = label_encoder.fit_transform(df.sex)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "y = df.survived.as_matrix()\n",
      "del df['survived']"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "X = df.as_matrix().astype(float)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "X_train, X_test, y_train, y_test = cross_validation.train_test_split(X, y, test_size=.5,\n",
      "                                                    random_state=2)\n",
      "\n",
      "classifier_name_tuples = [('Sigmoid Kernel Support Vector Classifier', svm.SVC(kernel='sigmoid', probability=True, shrinking=True, gamma=.1)),\n",
      "                          ('Bernoulli Naive Bayes Classifier', naive_bayes.BernoulliNB()),\n",
      "                          ('RBF Kernel Support Vector Classifier', svm.SVC(kernel='rbf', probability=True)),\n",
      "                          ('Extra Trees Classifier', ensemble.ExtraTreesClassifier()),\n",
      "                          ('Logistic Regression', linear_model.LogisticRegression(penalty='l1', C=.1)),\n",
      "                          ('Random Forest Classifier', ensemble.RandomForestClassifier()),\n",
      "                          ('Multinomial Naive Bayes Classifier', naive_bayes.MultinomialNB())\n",
      "                          ]\n",
      "\n",
      "    \n",
      "colors = itertools.cycle([\"#30a2da\",\n",
      "    \"#fc4f30\",\n",
      "    \"#e5ae38\",\n",
      "    \"#6d904f\",\n",
      "    \"#8b8b8b\",\n",
      "    \"dodgerblue\",\n",
      "    \"mediumpurple\",\n",
      "])\n",
      "\n",
      "fig, ax = plt.subplots(figsize=(16,12))\n",
      "for clf_name, clf in classifier_name_tuples:\n",
      "    clf.fit(X_train, y_train)\n",
      "    probas_ = clf.predict_proba(X_test)\n",
      "    fpr, tpr, thresholds = metrics.roc_curve(y_test, probas_[:, 1], )\n",
      "    roc_auc = metrics.auc(fpr, tpr)\n",
      "    plt.plot(fpr, tpr, color=colors.next(), lw=2, label='ROC curve for {}'.format(clf_name))\n",
      "    print('Region under curve for {0} is : {1:.2f}'.format(clf_name, roc_auc))\n",
      "#    ax.annotate('AUC for {0} = {1:.2f}'.format(clf_name, roc_auc), xy=(0.35, 0.6));\n",
      "\n",
      "plt.plot([0, 1], [0, 1], color='black', lw=2, linestyle='dotted', label='random guess')\n",
      "plt.xlim([0.0, 1.0])\n",
      "plt.ylim([0.0, 1.0])\n",
      "plt.xlabel('False Positive Rate');\n",
      "plt.ylabel('True Positive Rate');\n",
      "plt.title('ROC curves for different classifiers');\n",
      "plt.legend(loc=\"lower right\");"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}