finding_donors.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Machine Learning Engineer Nanodegree\n",
    "## Supervised Learning\n",
    "## Project: Finding Donors for *CharityML*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Welcome to the second project of the Machine Learning Engineer Nanodegree! In this notebook, some template code has already been provided for you, and it will be your job to implement the additional functionality necessary to successfully complete this project. Sections that begin with **'Implementation'** in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section and the specifics of the implementation are marked in the code block with a `'TODO'` statement. Please be sure to read the instructions carefully!\n",
    "\n",
    "In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a **'Question X'** header. Carefully read each question and provide thorough answers in the following text boxes that begin with **'Answer:'**. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.  \n",
    "\n",
    ">**Note:** Please specify WHICH VERSION OF PYTHON you are using when submitting this notebook. Code and Markdown cells can be executed using the **Shift + Enter** keyboard shortcut. In addition, Markdown cells can be edited by typically double-clicking the cell to enter edit mode."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Getting Started\n",
    "\n",
    "In this project, you will employ several supervised algorithms of your choice to accurately model individuals' income using data collected from the 1994 U.S. Census. You will then choose the best candidate algorithm from preliminary results and further optimize this algorithm to best model the data. Your goal with this implementation is to construct a model that accurately predicts whether an individual makes more than $50,000. This sort of task can arise in a non-profit setting, where organizations survive on donations.  Understanding an individual's income can help a non-profit better understand how large of a donation to request, or whether or not they should reach out to begin with.  While it can be difficult to determine an individual's general income bracket directly from public sources, we can (as we will see) infer this value from other publically available features. \n",
    "\n",
    "The dataset for this project originates from the [UCI Machine Learning Repository](https://archive.ics.uci.edu/ml/datasets/Census+Income). The datset was donated by Ron Kohavi and Barry Becker, after being published in the article _\"Scaling Up the Accuracy of Naive-Bayes Classifiers: A Decision-Tree Hybrid\"_. You can find the article by Ron Kohavi [online](https://www.aaai.org/Papers/KDD/1996/KDD96-033.pdf). The data we investigate here consists of small changes to the original dataset, such as removing the `'fnlwgt'` feature and records with missing or ill-formatted entries."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n",
    "## Exploring the Data\n",
    "Run the code cell below to load necessary Python libraries and load the census data. Note that the last column from this dataset, `'income'`, will be our target label (whether an individual makes more than, or at most, $50,000 annually). All other columns are features about each individual in the census database."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education_level</th>\n",
       "      <th>education-num</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>capital-gain</th>\n",
       "      <th>capital-loss</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>income</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13.0</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>2174.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age   workclass education_level  education-num  marital-status  \\\n",
       "0   39   State-gov       Bachelors           13.0   Never-married   \n",
       "\n",
       "      occupation    relationship    race    sex  capital-gain  capital-loss  \\\n",
       "0   Adm-clerical   Not-in-family   White   Male        2174.0           0.0   \n",
       "\n",
       "   hours-per-week  native-country income  \n",
       "0            40.0   United-States  <=50K  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Import libraries necessary for this project\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from time import time\n",
    "from IPython.display import display # Allows the use of display() for DataFrames\n",
    "\n",
    "# Import supplementary visualization code visuals.py\n",
    "import visuals as vs\n",
    "\n",
    "# Pretty display for notebooks\n",
    "%matplotlib inline\n",
    "\n",
    "# Load the Census dataset\n",
    "data = pd.read_csv(\"census.csv\")\n",
    "\n",
    "# Success - Display the first record\n",
    "display(data.head(n=1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementation: Data Exploration\n",
    "A cursory investigation of the dataset will determine how many individuals fit into either group, and will tell us about the percentage of these individuals making more than \\$50,000. In the code cell below, you will need to compute the following:\n",
    "- The total number of records, `'n_records'`\n",
    "- The number of individuals making more than \\$50,000 annually, `'n_greater_50k'`.\n",
    "- The number of individuals making at most \\$50,000 annually, `'n_at_most_50k'`.\n",
    "- The percentage of individuals making more than \\$50,000 annually, `'greater_percent'`.\n",
    "\n",
    "** HINT: ** You may need to look at the table above to understand how the `'income'` entries are formatted. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of records: 45222\n",
      "Individuals making more than $50,000: 11208\n",
      "Individuals making at most $50,000: 34014\n",
      "Percentage of individuals making more than $50,000: 24.78%\n"
     ]
    }
   ],
   "source": [
    "# TODO: Total number of records\n",
    "n_records = len(data.index)\n",
    "\n",
    "# TODO: Number of records where individual's income is more than $50,000\n",
    "n_greater_50k = len(data[data.income == '>50K'])\n",
    "\n",
    "# TODO: Number of records where individual's income is at most $50,000\n",
    "n_at_most_50k = len(data[data.income == '<=50K'])\n",
    "\n",
    "# TODO: Percentage of individuals whose income is more than $50,000\n",
    "greater_percent = (n_greater_50k / float(n_records)) * 100\n",
    "\n",
    "# Print the results\n",
    "print(\"Total number of records: {}\".format(n_records))\n",
    "print(\"Individuals making more than $50,000: {}\".format(n_greater_50k))\n",
    "print(\"Individuals making at most $50,000: {}\".format(n_at_most_50k))\n",
    "print(\"Percentage of individuals making more than $50,000: {:.2f}%\".format(greater_percent))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "** Featureset Exploration **\n",
    "\n",
    "* **age**: continuous. \n",
    "* **workclass**: Private, Self-emp-not-inc, Self-emp-inc, Federal-gov, Local-gov, State-gov, Without-pay, Never-worked. \n",
    "* **education**: Bachelors, Some-college, 11th, HS-grad, Prof-school, Assoc-acdm, Assoc-voc, 9th, 7th-8th, 12th, Masters, 1st-4th, 10th, Doctorate, 5th-6th, Preschool. \n",
    "* **education-num**: continuous. \n",
    "* **marital-status**: Married-civ-spouse, Divorced, Never-married, Separated, Widowed, Married-spouse-absent, Married-AF-spouse. \n",
    "* **occupation**: Tech-support, Craft-repair, Other-service, Sales, Exec-managerial, Prof-specialty, Handlers-cleaners, Machine-op-inspct, Adm-clerical, Farming-fishing, Transport-moving, Priv-house-serv, Protective-serv, Armed-Forces. \n",
    "* **relationship**: Wife, Own-child, Husband, Not-in-family, Other-relative, Unmarried. \n",
    "* **race**: Black, White, Asian-Pac-Islander, Amer-Indian-Eskimo, Other. \n",
    "* **sex**: Female, Male. \n",
    "* **capital-gain**: continuous. \n",
    "* **capital-loss**: continuous. \n",
    "* **hours-per-week**: continuous. \n",
    "* **native-country**: United-States, Cambodia, England, Puerto-Rico, Canada, Germany, Outlying-US(Guam-USVI-etc), India, Japan, Greece, South, China, Cuba, Iran, Honduras, Philippines, Italy, Poland, Jamaica, Vietnam, Mexico, Portugal, Ireland, France, Dominican-Republic, Laos, Ecuador, Taiwan, Haiti, Columbia, Hungary, Guatemala, Nicaragua, Scotland, Thailand, Yugoslavia, El-Salvador, Trinadad&Tobago, Peru, Hong, Holand-Netherlands."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n",
    "## Preparing the Data\n",
    "Before data can be used as input for machine learning algorithms, it often must be cleaned, formatted, and restructured — this is typically known as **preprocessing**. Fortunately, for this dataset, there are no invalid or missing entries we must deal with, however, there are some qualities about certain features that must be adjusted. This preprocessing can help tremendously with the outcome and predictive power of nearly all learning algorithms."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Transforming Skewed Continuous Features\n",
    "A dataset may sometimes contain at least one feature whose values tend to lie near a single number, but will also have a non-trivial number of vastly larger or smaller values than that single number.  Algorithms can be sensitive to such distributions of values and can underperform if the range is not properly normalized. With the census dataset two features fit this description: '`capital-gain'` and `'capital-loss'`. \n",
    "\n",
    "Run the code cell below to plot a histogram of these two features. Note the range of the values present and how they are distributed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxAAAAF2CAYAAAD+y36TAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n2gQEAADQJiAAAIA2AQEAALQJCAAAoO0Ga7UEP7pgqRcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10d49c290>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Split the data into features and target label\n",
    "income_raw = data['income']\n",
    "features_raw = data.drop('income', axis = 1)\n",
    "\n",
    "# Visualize skewed continuous features of original data\n",
    "vs.distribution(data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For highly-skewed feature distributions such as `'capital-gain'` and `'capital-loss'`, it is common practice to apply a <a href=\"https://en.wikipedia.org/wiki/Data_transformation_(statistics)\">logarithmic transformation</a> on the data so that the very large and very small values do not negatively affect the performance of a learning algorithm. Using a logarithmic transformation significantly reduces the range of values caused by outliers. Care must be taken when applying this transformation however: The logarithm of `0` is undefined, so we must translate the values by a small amount above `0` to apply the the logarithm successfully.\n",
    "\n",
    "Run the code cell below to perform a transformation on the data and visualize the results. Again, note the range of values and how they are distributed. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxAAAAF2CAYAAAD+y36TAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlt95AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10dd8a8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Log-transform the skewed features\n",
    "skewed = ['capital-gain', 'capital-loss']\n",
    "features_log_transformed = pd.DataFrame(data = features_raw)\n",
    "features_log_transformed[skewed] = features_raw[skewed].apply(lambda x: np.log(x + 1))\n",
    "\n",
    "# Visualize the new log distributions\n",
    "vs.distribution(features_log_transformed, transformed = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Normalizing Numerical Features\n",
    "In addition to performing transformations on features that are highly skewed, it is often good practice to perform some type of scaling on numerical features. Applying a scaling to the data does not change the shape of each feature's distribution (such as `'capital-gain'` or `'capital-loss'` above); however, normalization ensures that each feature is treated equally when applying supervised learners. Note that once scaling is applied, observing the data in its raw form will no longer have the same original meaning, as exampled below.\n",
    "\n",
    "Run the code cell below to normalize each numerical feature. We will use [`sklearn.preprocessing.MinMaxScaler`](http://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.MinMaxScaler.html) for this."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education_level</th>\n",
       "      <th>education-num</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>capital-gain</th>\n",
       "      <th>capital-loss</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.301370</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.667492</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.397959</td>\n",
       "      <td>United-States</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.452055</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.122449</td>\n",
       "      <td>United-States</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.287671</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>Divorced</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.397959</td>\n",
       "      <td>United-States</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.493151</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.397959</td>\n",
       "      <td>United-States</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.150685</td>\n",
       "      <td>Private</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>0.800000</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>Wife</td>\n",
       "      <td>Black</td>\n",
       "      <td>Female</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.397959</td>\n",
       "      <td>Cuba</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        age          workclass education_level  education-num  \\\n",
       "0  0.301370          State-gov       Bachelors       0.800000   \n",
       "1  0.452055   Self-emp-not-inc       Bachelors       0.800000   \n",
       "2  0.287671            Private         HS-grad       0.533333   \n",
       "3  0.493151            Private            11th       0.400000   \n",
       "4  0.150685            Private       Bachelors       0.800000   \n",
       "\n",
       "        marital-status          occupation    relationship    race      sex  \\\n",
       "0        Never-married        Adm-clerical   Not-in-family   White     Male   \n",
       "1   Married-civ-spouse     Exec-managerial         Husband   White     Male   \n",
       "2             Divorced   Handlers-cleaners   Not-in-family   White     Male   \n",
       "3   Married-civ-spouse   Handlers-cleaners         Husband   Black     Male   \n",
       "4   Married-civ-spouse      Prof-specialty            Wife   Black   Female   \n",
       "\n",
       "   capital-gain  capital-loss  hours-per-week  native-country  \n",
       "0      0.667492           0.0        0.397959   United-States  \n",
       "1      0.000000           0.0        0.122449   United-States  \n",
       "2      0.000000           0.0        0.397959   United-States  \n",
       "3      0.000000           0.0        0.397959   United-States  \n",
       "4      0.000000           0.0        0.397959            Cuba  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Import sklearn.preprocessing.StandardScaler\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "# Initialize a scaler, then apply it to the features\n",
    "scaler = MinMaxScaler() # default=(0, 1)\n",
    "numerical = ['age', 'education-num', 'capital-gain', 'capital-loss', 'hours-per-week']\n",
    "\n",
    "features_log_minmax_transform = pd.DataFrame(data = features_log_transformed)\n",
    "features_log_minmax_transform[numerical] = scaler.fit_transform(features_log_transformed[numerical])\n",
    "\n",
    "# Show an example of a record with scaling applied\n",
    "display(features_log_minmax_transform.head(n = 5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementation: Data Preprocessing\n",
    "\n",
    "From the table in **Exploring the Data** above, we can see there are several features for each record that are non-numeric. Typically, learning algorithms expect input to be numeric, which requires that non-numeric features (called *categorical variables*) be converted. One popular way to convert categorical variables is by using the **one-hot encoding** scheme. One-hot encoding creates a _\"dummy\"_ variable for each possible category of each non-numeric feature. For example, assume `someFeature` has three possible entries: `A`, `B`, or `C`. We then encode this feature into `someFeature_A`, `someFeature_B` and `someFeature_C`.\n",
    "\n",
    "|   | someFeature |                    | someFeature_A | someFeature_B | someFeature_C |\n",
    "| :-: | :-: |                            | :-: | :-: | :-: |\n",
    "| 0 |  B  |  | 0 | 1 | 0 |\n",
    "| 1 |  C  | ----> one-hot encode ----> | 0 | 0 | 1 |\n",
    "| 2 |  A  |  | 1 | 0 | 0 |\n",
    "\n",
    "Additionally, as with the non-numeric features, we need to convert the non-numeric target label, `'income'` to numerical values for the learning algorithm to work. Since there are only two possible categories for this label (\"<=50K\" and \">50K\"), we can avoid using one-hot encoding and simply encode these two categories as `0` and `1`, respectively. In code cell below, you will need to implement the following:\n",
    " - Use [`pandas.get_dummies()`](http://pandas.pydata.org/pandas-docs/stable/generated/pandas.get_dummies.html?highlight=get_dummies#pandas.get_dummies) to perform one-hot encoding on the `'features_raw'` data.\n",
    " - Convert the target label `'income_raw'` to numerical entries.\n",
    "   - Set records with \"<=50K\" to `0` and records with \">50K\" to `1`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "103 total features after one-hot encoding.\n",
      "['age', 'education-num', 'capital-gain', 'capital-loss', 'hours-per-week', 'workclass_ Federal-gov', 'workclass_ Local-gov', 'workclass_ Private', 'workclass_ Self-emp-inc', 'workclass_ Self-emp-not-inc', 'workclass_ State-gov', 'workclass_ Without-pay', 'education_level_ 10th', 'education_level_ 11th', 'education_level_ 12th', 'education_level_ 1st-4th', 'education_level_ 5th-6th', 'education_level_ 7th-8th', 'education_level_ 9th', 'education_level_ Assoc-acdm', 'education_level_ Assoc-voc', 'education_level_ Bachelors', 'education_level_ Doctorate', 'education_level_ HS-grad', 'education_level_ Masters', 'education_level_ Preschool', 'education_level_ Prof-school', 'education_level_ Some-college', 'marital-status_ Divorced', 'marital-status_ Married-AF-spouse', 'marital-status_ Married-civ-spouse', 'marital-status_ Married-spouse-absent', 'marital-status_ Never-married', 'marital-status_ Separated', 'marital-status_ Widowed', 'occupation_ Adm-clerical', 'occupation_ Armed-Forces', 'occupation_ Craft-repair', 'occupation_ Exec-managerial', 'occupation_ Farming-fishing', 'occupation_ Handlers-cleaners', 'occupation_ Machine-op-inspct', 'occupation_ Other-service', 'occupation_ Priv-house-serv', 'occupation_ Prof-specialty', 'occupation_ Protective-serv', 'occupation_ Sales', 'occupation_ Tech-support', 'occupation_ Transport-moving', 'relationship_ Husband', 'relationship_ Not-in-family', 'relationship_ Other-relative', 'relationship_ Own-child', 'relationship_ Unmarried', 'relationship_ Wife', 'race_ Amer-Indian-Eskimo', 'race_ Asian-Pac-Islander', 'race_ Black', 'race_ Other', 'race_ White', 'sex_ Female', 'sex_ Male', 'native-country_ Cambodia', 'native-country_ Canada', 'native-country_ China', 'native-country_ Columbia', 'native-country_ Cuba', 'native-country_ Dominican-Republic', 'native-country_ Ecuador', 'native-country_ El-Salvador', 'native-country_ England', 'native-country_ France', 'native-country_ Germany', 'native-country_ Greece', 'native-country_ Guatemala', 'native-country_ Haiti', 'native-country_ Holand-Netherlands', 'native-country_ Honduras', 'native-country_ Hong', 'native-country_ Hungary', 'native-country_ India', 'native-country_ Iran', 'native-country_ Ireland', 'native-country_ Italy', 'native-country_ Jamaica', 'native-country_ Japan', 'native-country_ Laos', 'native-country_ Mexico', 'native-country_ Nicaragua', 'native-country_ Outlying-US(Guam-USVI-etc)', 'native-country_ Peru', 'native-country_ Philippines', 'native-country_ Poland', 'native-country_ Portugal', 'native-country_ Puerto-Rico', 'native-country_ Scotland', 'native-country_ South', 'native-country_ Taiwan', 'native-country_ Thailand', 'native-country_ Trinadad&Tobago', 'native-country_ United-States', 'native-country_ Vietnam', 'native-country_ Yugoslavia']\n"
     ]
    }
   ],
   "source": [
    "# TODO: One-hot encode the 'features_log_minmax_transform' data using pandas.get_dummies()\n",
    "features_final = pd.get_dummies(features_log_minmax_transform)\n",
    "\n",
    "# TODO: Encode the 'income_raw' data to numerical values\n",
    "income = income_raw.replace({'<=50K':0, '>50K':1})\n",
    "\n",
    "# Print the number of features after one-hot encoding\n",
    "encoded = list(features_final.columns)\n",
    "print(\"{} total features after one-hot encoding.\".format(len(encoded)))\n",
    "\n",
    "# Uncomment the following line to see the encoded feature names\n",
    "print(encoded)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Shuffle and Split Data\n",
    "Now all _categorical variables_ have been converted into numerical features, and all numerical features have been normalized. As always, we will now split the data (both features and their labels) into training and test sets. 80% of the data will be used for training and 20% for testing.\n",
    "\n",
    "Run the code cell below to perform this split."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set has 36177 samples.\n",
      "Testing set has 9045 samples.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/robertyoung/anaconda/lib/python2.7/site-packages/sklearn/cross_validation.py:44: DeprecationWarning: This module was deprecated in version 0.18 in favor of the model_selection module into which all the refactored classes and functions are moved. Also note that the interface of the new CV iterators are different from that of this module. This module will be removed in 0.20.\n",
      "  \"This module will be removed in 0.20.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "# Import train_test_split\n",
    "from sklearn.cross_validation import train_test_split\n",
    "\n",
    "# Split the 'features' and 'income' data into training and testing sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(features_final, \n",
    "                                                    income, \n",
    "                                                    test_size = 0.2, \n",
    "                                                    random_state = 0)\n",
    "\n",
    "# Show the results of the split\n",
    "print(\"Training set has {} samples.\".format(X_train.shape[0]))\n",
    "print(\"Testing set has {} samples.\".format(X_test.shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n",
    "## Evaluating Model Performance\n",
    "In this section, we will investigate four different algorithms, and determine which is best at modeling the data. Three of these algorithms will be supervised learners of your choice, and the fourth algorithm is known as a *naive predictor*."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Metrics and the Naive Predictor\n",
    "*CharityML*, equipped with their research, knows individuals that make more than \\$50,000 are most likely to donate to their charity. Because of this, *CharityML* is particularly interested in predicting who makes more than \\$50,000 accurately. It would seem that using **accuracy** as a metric for evaluating a particular model's performace would be appropriate. Additionally, identifying someone that *does not* make more than \\$50,000 as someone who does would be detrimental to *CharityML*, since they are looking to find individuals willing to donate. Therefore, a model's ability to precisely predict those that make more than \\$50,000 is *more important* than the model's ability to **recall** those individuals. We can use **F-beta score** as a metric that considers both precision and recall:\n",
    "\n",
    "$$ F_{\\beta} = (1 + \\beta^2) \\cdot \\frac{precision \\cdot recall}{\\left( \\beta^2 \\cdot precision \\right) + recall} $$\n",
    "\n",
    "In particular, when $\\beta = 0.5$, more emphasis is placed on precision. This is called the **F$_{0.5}$ score** (or F-score for simplicity).\n",
    "\n",
    "Looking at the distribution of classes (those who make at most \\$50,000, and those who make more), it's clear most individuals do not make more than \\$50,000. This can greatly affect **accuracy**, since we could simply say *\"this person does not make more than \\$50,000\"* and generally be right, without ever looking at the data! Making such a statement would be called **naive**, since we have not considered any information to substantiate the claim. It is always important to consider the *naive prediction* for your data, to help establish a benchmark for whether a model is performing well. That been said, using that prediction would be pointless: If we predicted all people made less than \\$50,000, *CharityML* would identify no one as donors. \n",
    "\n",
    "\n",
    "#### Note: Recap of accuracy, precision, recall\n",
    "\n",
    "** Accuracy ** measures how often the classifier makes the correct prediction. It’s the ratio of the number of correct predictions to the total number of predictions (the number of test data points).\n",
    "\n",
    "** Precision ** tells us what proportion of messages we classified as spam, actually were spam.\n",
    "It is a ratio of true positives(words classified as spam, and which are actually spam) to all positives(all words classified as spam, irrespective of whether that was the correct classificatio), in other words it is the ratio of\n",
    "\n",
    "`[True Positives/(True Positives + False Positives)]`\n",
    "\n",
    "** Recall(sensitivity)** tells us what proportion of messages that actually were spam were classified by us as spam.\n",
    "It is a ratio of true positives(words classified as spam, and which are actually spam) to all the words that were actually spam, in other words it is the ratio of\n",
    "\n",
    "`[True Positives/(True Positives + False Negatives)]`\n",
    "\n",
    "For classification problems that are skewed in their classification distributions like in our case, for example if we had a 100 text messages and only 2 were spam and the rest 98 weren't, accuracy by itself is not a very good metric. We could classify 90 messages as not spam(including the 2 that were spam but we classify them as not spam, hence they would be false negatives) and 10 as spam(all 10 false positives) and still get a reasonably good accuracy score. For such cases, precision and recall come in very handy. These two metrics can be combined to get the F1 score, which is weighted average(harmonic mean) of the precision and recall scores. This score can range from 0 to 1, with 1 being the best possible F1 score(we take the harmonic mean as we are dealing with ratios)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 1 - Naive Predictor Performance\n",
    "* If we chose a model that always predicted an individual made more than $50,000, what would  that model's accuracy and F-score be on this dataset? You must use the code cell below and assign your results to `'accuracy'` and `'fscore'` to be used later.\n",
    "\n",
    "** HINT: ** \n",
    "\n",
    "* When we have a model that always predicts '1' (i.e. the individual makes more than 50k) then our model will have no True Negatives(TN) or False Negatives(FN) as we are not making any negative('0' value) predictions. Therefore our Accuracy in this case becomes the same as our Precision(True Positives/(True Positives + False Positives)) as every prediction that we have made with value '1' that should have '0' becomes a False Positive; therefore our denominator in this case is the total number of records we have in total. \n",
    "* Our Recall score(True Positives/(True Positives + False Negatives)) in this setting becomes 1 as we have no False Negatives."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Naive Predictor: [Accuracy score: 0.2478, F-score: 0.2917]\n"
     ]
    }
   ],
   "source": [
    "'''\n",
    "TP = np.sum(income) # Counting the ones as this is the naive case. Note that 'income' is the 'income_raw' data \n",
    "encoded to numerical values done in the data preprocessing step.\n",
    "FP = income.count() - TP # Specific to the naive case\n",
    "\n",
    "TN = 0 # No predicted negatives in the naive case\n",
    "FN = 0 # No predicted negatives in the naive case\n",
    "'''\n",
    "# TODO: Calculate accuracy, precision and recall\n",
    "accuracy = (np.sum(income)) / ((np.sum(income)) + float(((income.count()) - np.sum(income))))\n",
    "recall = np.sum(income) / float((np.sum(income) + 0))\n",
    "precision = np.sum(income) / float(((np.sum(income) + ((income.count()) - np.sum(income)))))\n",
    "beta = 0.5\n",
    "\n",
    "# TODO: Calculate F-score using the formula above for beta = 0.5 and correct values for precision and recall.\n",
    "# HINT: The formula above can be written as (1 + beta**2) * (precision * recall) / ((beta**2 * precision) + recall)\n",
    "fscore = (1+ beta**2) * (precision * recall) / ((beta ** 2 * precision) + recall)\n",
    "\n",
    "# Print the results \n",
    "\n",
    "print(\"Naive Predictor: [Accuracy score: {:.4f}, F-score: {:.4f}]\".format(accuracy, fscore))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Supervised Learning Models\n",
    "**The following are some of the supervised learning models that are currently available in** [`scikit-learn`](http://scikit-learn.org/stable/supervised_learning.html) **that you may choose from:**\n",
    "- Gaussian Naive Bayes (GaussianNB)\n",
    "- Decision Trees\n",
    "- Ensemble Methods (Bagging, AdaBoost, Random Forest, Gradient Boosting)\n",
    "- K-Nearest Neighbors (KNeighbors)\n",
    "- Stochastic Gradient Descent Classifier (SGDC)\n",
    "- Support Vector Machines (SVM)\n",
    "- Logistic Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 2 - Model Application\n",
    "List three of the supervised learning models above that are appropriate for this problem that you will test on the census data. For each model chosen\n",
    "\n",
    "- Describe one real-world application in industry where the model can be applied. \n",
    "- What are the strengths of the model; when does it perform well?\n",
    "- What are the weaknesses of the model; when does it perform poorly?\n",
    "- What makes this model a good candidate for the problem, given what you know about the data?\n",
    "\n",
    "** HINT: **\n",
    "\n",
    "Structure your answer in the same format as above^, with 4 parts for each of the three models you pick. Please include references with your answer."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer: **\n",
    "\n",
    "**Support Vector Machines:**\n",
    "\n",
    "One real world application:\n",
    "* Hand writing recognition can be deployed using SVM. [1]\n",
    "\n",
    "Strengths of the model:\n",
    "* Works well in high dimension spaces - for example, image recognition where every pixel may be treated as a feature. [2]\n",
    "* Maintains effectiveness even in cases where the number of dimensions exceeds the number of samples - again applicable in the field of image recognition. [2]\n",
    "* It is memory efficient due to its use of a subset of training points in the decision function. [2]\n",
    "* It provides versatility through the deployment of common and custom kernels. [2]\n",
    "* Works well in a complicated domain where there is a clear degree of seperation. [3]\n",
    "\n",
    "\n",
    "Weaknesses of the model:\n",
    "* Overfitting must be avoided through the selection of the correct kernel choice and regularisation term *if* the number of features far exceeds the number of samples. [2]\n",
    "* Inefficient five-fold cross-validation is deployed to calculate probability estimates. [2]\n",
    "* Can only be applied to two class tasks. As a result, multi-class tasks must be reduced to several binary problems. \n",
    "* The parameters of a model that has been solved can be difficult to interpret. \n",
    "* Does not perform well in large datasets as training is cubic in the size of the dataset. [3]\n",
    "* Does not work well with lots of noise, so when classes are overlapping you have to count independant evidence (this is where a NB classifier would work better). [3]\n",
    "\n",
    "\n",
    "What makes it a good candidate for the problem [6]:\n",
    "* Sample size is greater than 50 samples (have enough data to train with).\n",
    "* Data is labelled.\n",
    "* Predicting a categorey (works with classification)\n",
    "* Sample size is less than 100k.\n",
    "* Works well with large feature sets (ours is relatively small -> medium size).\n",
    "\n",
    "\n",
    "\n",
    "**SGD Classifier:**\n",
    "\n",
    "One real world application:\n",
    "* Text classification and natural language processing. It is useful as when the given data is sparse, the function can easily scale to problems with more than 10^5 training examples and more than 10^5 features. [4]\n",
    "\n",
    "Strengths of the model:\n",
    "* It is efficient. [4] \n",
    "* It is easy to implement and provides a lot of opportunities for code tuning. [4]\n",
    "\n",
    "Weaknesses of the model:\n",
    "* A number of hyperparameters are required for SGD, such as the number of iterations and the regularisation parameter. [4]\n",
    "* It (SGD) is sensitive to feature scaling. [4]\n",
    "\n",
    "What makes it a good candidate for the problem [6]:\n",
    "* Sample size is greater than 50 samples (have enough data to train with).\n",
    "* Data is labelled.\n",
    "* Predicting a categorey (works with classification)\n",
    "* Sample size is less than 100k, but SGD works well with sizes greater than this, meaning greater computational efficiency and speed. \n",
    "\n",
    "\n",
    "**K Nearest Neighbour:**\n",
    "\n",
    "One real world application:\n",
    "* KNN can be used to provide recommendations. A real world example of this would be video streaming services such as Netflix or Amazon Prime. If a given user likes an item in the library, similar items that they may like, but are unaware of can be recommended to them by using data from other users and their likes. If it is seen that a similar set of users like two different items, these items are probably similar and to each the respective users taste, and worthy of a recommendation. \n",
    "\n",
    "Strengths of the model:\n",
    "* Easy to understand and implement - not much code is required. [5]\n",
    "* No probability distributions are assumed based on the input data. This is useful with inputs where the probability distribution is unknown, making it robust. [5]\n",
    "* KNN is a lazy learner. This means it generalises data during the training phase, not the testing phase. This allows it quickly adapt to changes as it does not expect a generalised data set. [5]\n",
    "\n",
    "Weaknesses of the model:\n",
    "* KNN gets its information from its input neighbours. As a result of this, localised outliers can affect outcomes significantly when compared with other algorithms which have a generalised view of the data. It is sensitive to localised data. [5]\n",
    "* One of its strengths, lazy-learning, is also one of its weaknesses. As most of the computation is done during testing, rather than during training, this can result in long computation times when dealing with large datasets. [5]\n",
    "* If there is a type of categorey that is present much more than another, classifying an input will result in a bias to this more abundent categorey. This can be dealt with by adjusting the weights based on occurences, but will still pose a problem near the decision boundary. [5]\n",
    "* Inputs can be close to many points when there are many dimensions. The effectiveness of k-NN is reduced as a result. This is as it relies on the correlation between closeness and similarity. Dimension reduction can be used to reduce the effects of this, but variable trends may be lost as a result. [5]\n",
    "\n",
    "What makes it a good candidate for the problem [6]:\n",
    "* Sample size is greater than 50 samples (have enough data to train with).\n",
    "* Predicting a categorey (works with classification).\n",
    "* Sample size is less than 100k\n",
    "* Data is labelled.\n",
    "\n",
    "\n",
    "**References (denoted through the use of square brackets and numbering):**\n",
    "\n",
    "1. https://en.wikipedia.org/wiki/Support_vector_machine#Applications\n",
    "2. http://scikit-learn.org/stable/modules/svm.html\n",
    "3. Udacity lectures (Machine Learning Nano Degree)\n",
    "4. http://scikit-learn.org/stable/modules/sgd.html\n",
    "5. https://brilliant.org/wiki/k-nearest-neighbors/#pros-and-cons\n",
    "6. http://scikit-learn.org/stable/tutorial/machine_learning_map/index.html"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementation - Creating a Training and Predicting Pipeline\n",
    "To properly evaluate the performance of each model you've chosen, it's important that you create a training and predicting pipeline that allows you to quickly and effectively train models using various sizes of training data and perform predictions on the testing data. Your implementation here will be used in the following section.\n",
    "In the code block below, you will need to implement the following:\n",
    " - Import `fbeta_score` and `accuracy_score` from [`sklearn.metrics`](http://scikit-learn.org/stable/modules/classes.html#sklearn-metrics-metrics).\n",
    " - Fit the learner to the sampled training data and record the training time.\n",
    " - Perform predictions on the test data `X_test`, and also on the first 300 training points `X_train[:300]`.\n",
    "   - Record the total prediction time.\n",
    " - Calculate the accuracy score for both the training subset and testing set.\n",
    " - Calculate the F-score for both the training subset and testing set.\n",
    "   - Make sure that you set the `beta` parameter!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# TODO: Import two metrics from sklearn - fbeta_score and accuracy_score\n",
    "\n",
    "from sklearn.metrics import fbeta_score, accuracy_score\n",
    "\n",
    "\n",
    "def train_predict(learner, sample_size, X_train, y_train, X_test, y_test): \n",
    "    '''\n",
    "    inputs:\n",
    "       - learner: the learning algorithm to be trained and predicted on\n",
    "       - sample_size: the size of samples (number) to be drawn from training set\n",
    "       - X_train: features training set\n",
    "       - y_train: income training set\n",
    "       - X_test: features testing set\n",
    "       - y_test: income testing set\n",
    "    '''\n",
    "    \n",
    "    results = {}\n",
    "    \n",
    "    # TODO: Fit the learner to the training data using slicing with 'sample_size' using .fit(training_features[:], training_labels[:])\n",
    "    start = time() # Get start time\n",
    "    learner = learner.fit(X_train[:sample_size], y_train[:sample_size])\n",
    "    end = time() # Get end time\n",
    "    \n",
    "    # TODO: Calculate the training time\n",
    "    results['train_time'] = end - start\n",
    "        \n",
    "    # TODO: Get the predictions on the test set(X_test),\n",
    "    #       then get predictions on the first 300 training samples(X_train) using .predict()\n",
    "    start = time() # Get start time\n",
    "    predictions_test = learner.predict(X_test)\n",
    "    predictions_train = learner.predict(X_train[:300])\n",
    "    end = time() # Get end times\n",
    "    \n",
    "    # TODO: Calculate the total prediction time\n",
    "    results['pred_time'] = end - start\n",
    "            \n",
    "    # TODO: Compute accuracy on the first 300 training samples which is y_train[:300]\n",
    "    results['acc_train'] = accuracy_score(y_train[:300], predictions_train)\n",
    "        \n",
    "    # TODO: Compute accuracy on test set using accuracy_score()\n",
    "    results['acc_test'] = accuracy_score(y_test, predictions_test)\n",
    "    \n",
    "    # TODO: Compute F-score on the the first 300 training samples using fbeta_score()\n",
    "    results['f_train'] = fbeta_score(y_train[:300], predictions_train[:300], beta = 0.5)\n",
    "        \n",
    "    # TODO: Compute F-score on the test set which is y_test\n",
    "    results['f_test'] = fbeta_score(y_test, predictions_test, beta = 0.5)\n",
    "       \n",
    "    # Success\n",
    "    print(\"{} trained on {} samples.\".format(learner.__class__.__name__, sample_size))\n",
    "        \n",
    "    # Return the results\n",
    "    return results\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementation: Initial Model Evaluation\n",
    "In the code cell, you will need to implement the following:\n",
    "- Import the three supervised learning models you've discussed in the previous section.\n",
    "- Initialize the three models and store them in `'clf_A'`, `'clf_B'`, and `'clf_C'`.\n",
    "  - Use a `'random_state'` for each model you use, if provided.\n",
    "  - **Note:** Use the default settings for each model — you will tune one specific model in a later section.\n",
    "- Calculate the number of records equal to 1%, 10%, and 100% of the training data.\n",
    "  - Store those values in `'samples_1'`, `'samples_10'`, and `'samples_100'` respectively.\n",
    "\n",
    "**Note:** Depending on which algorithms you chose, the following implementation may take some time to run!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/robertyoung/anaconda/lib/python2.7/site-packages/sklearn/metrics/classification.py:1113: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SVC trained on 361 samples.\n",
      "SVC trained on 3617 samples.\n",
      "SVC trained on 36177 samples.\n",
      "SGDClassifier trained on 361 samples.\n",
      "SGDClassifier trained on 3617 samples.\n",
      "SGDClassifier trained on 36177 samples.\n",
      "KNeighborsClassifier trained on 361 samples.\n",
      "KNeighborsClassifier trained on 3617 samples.\n",
      "KNeighborsClassifier trained on 36177 samples.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxAAAAIuCAYAAAAv/u6UAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x110c36a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# TODO: Import the three supervised learning models from sklearn\n",
    "from sklearn import svm\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "\n",
    "# TODO: Initialize the three models\n",
    "clf_A = svm.SVC(random_state = 7)\n",
    "clf_B = SGDClassifier(random_state = 7)\n",
    "clf_C = KNeighborsClassifier()\n",
    "\n",
    "\n",
    "\n",
    "# TODO: Calculate the number of samples for 1%, 10%, and 100% of the training data\n",
    "# HINT: samples_100 is the entire training set i.e. len(y_train)\n",
    "# HINT: samples_10 is 10% of samples_100\n",
    "# HINT: samples_1 is 1% of samples_100\n",
    "samples_100 = len(y_train)\n",
    "samples_10 = int(samples_100 * 0.1)\n",
    "samples_1 = int(samples_100 * 0.01)\n",
    "\n",
    "# Collect results on the learners\n",
    "results = {}\n",
    "for clf in [clf_A, clf_B, clf_C]:\n",
    "    clf_name = clf.__class__.__name__\n",
    "    results[clf_name] = {}\n",
    "    for i, samples in enumerate([samples_1, samples_10, samples_100]):\n",
    "        results[clf_name][i] = \\\n",
    "        train_predict(clf, samples, X_train, y_train, X_test, y_test)\n",
    "\n",
    "# Run metrics visualization for the three supervised learning models chosen\n",
    "vs.evaluate(results, accuracy, fscore)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n",
    "## Improving Results\n",
    "In this final section, you will choose from the three supervised learning models the *best* model to use on the student data. You will then perform a grid search optimization for the model over the entire training set (`X_train` and `y_train`) by tuning at least one parameter to improve upon the untuned model's F-score. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 3 - Choosing the Best Model\n",
    "\n",
    "* Based on the evaluation you performed earlier, in one to two paragraphs, explain to *CharityML* which of the three models you believe to be most appropriate for the task of identifying individuals that make more than \\$50,000. \n",
    "\n",
    "** HINT: ** \n",
    "Look at the graph at the bottom left from the cell above(the visualization created by `vs.evaluate(results, accuracy, fscore)`) and check the F score for the testing set when 100% of the training set is used. Which model has the highest score? Your answer should include discussion of the:\n",
    "* metrics - F score on the testing when 100% of the training data is used, \n",
    "* prediction/training time\n",
    "* the algorithm's suitability for the data.\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer: **\n",
    "\n",
    "Originally, for this answer, I had proposed SVC as being the best model to use for this application. However, on working through the assignment further, I found this not to be the case, and have instead opted for the SGD model. I will explain why in further detail below. \n",
    "\n",
    "F-Score:\n",
    "When 100% of the training set is used, the relative F-scores are as follows:\n",
    "* SVC - 0.674\n",
    "* SGD - 0.649\n",
    "* KNeighborsClassifier - 0.639\n",
    "\n",
    "SCV provides the best results with respect to F-score, achieving 67.4%. SGD is second best, delivering 64.9%, a delta of 2.5%. However, as will be discussed in the next section, F-score alone is not all that should be taken into consideration, as the prediction and training time is also of importance.\n",
    "\n",
    "Prediction/Training Time:\n",
    "The model training time for SVC is by far and away the longest, at just over 140 seconds. Both SGD and Kneighbours are so small that they barely register on the bar chart. \n",
    "The model predicting time is greatest for KNeighbours, due to its lazy learner characteristics. SVC follows this, at just under 25 seconds. \n",
    "Once again, SGD trumps all by having, by a large margin, the shortest model prediction time. And for my work modifying parameters in the later exercise, this was crucial. I found that SVC would take far too long to complete with each additional parameter tuning. In comparison, SGD, was done almost instantly, and only with the aforementioned accuracy drop of 2.5% in comparison.\n",
    "\n",
    "\n",
    "Algorithm Suitability:\n",
    "Due to its relatively low time taken to train and predict, and relatively high f-score, I would suggest using the SGD model. If time was not an issue, and accuracy was a priority for the Charity, SVC could instead be deployed.\n",
    "As mentioned in the previous section, the alorithim is suitable as:\n",
    "\n",
    "* Sample size is greater than 50 samples (have enough data to train with).\n",
    "* Data is labelled.\n",
    "* Predicting a categorey (works with classification)\n",
    "* It is efficient.\n",
    "* It is easy to implement and provides a lot of opportunities for code tuning. [4]\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 4 - Describing the Model in Layman's Terms\n",
    "\n",
    "* In one to two paragraphs, explain to *CharityML*, in layman's terms, how the final model chosen is supposed to work. Be sure that you are describing the major qualities of the model, such as how the model is trained and how the model makes a prediction. Avoid using advanced mathematical jargon, such as describing equations.\n",
    "\n",
    "** HINT: **\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer: ** \n",
    "\n",
    "**SGD [1]**\n",
    "\n",
    "In stochastic gradient descent, the model calculates the true gradient by approximating the gradient at a single example. At each training example, it will perform an update of this gradient based on these individual examples. This can happen until the algorithm reaches convergence. It attempts to find the minima or maxima by the process of iteration.\n",
    "\n",
    "Step by step, this will look as follows:\n",
    "\n",
    "1.\tAfter shuffle, split and initial training using the training data, an initial estimate of the gradient is made (based on the parameters and chosen learning rate). \n",
    "2.\tThe algorithm will go through the training set, in iterations, and at each training example , it will perform an adjustment on the gradient approximation. This is based on the feedback the current approximation gets versus the current training example.\n",
    "3.\tSeveral passes may be made on the training set until the algorithm begins to reach convergence (important to shuffle data to avoid cycles). \n",
    "    \n",
    "**References**\n",
    "\n",
    "1. https://en.wikipedia.org/wiki/Stochastic_gradient_descent#Background"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementation: Model Tuning\n",
    "Fine tune the chosen model. Use grid search (`GridSearchCV`) with at least one important parameter tuned with at least 3 different values. You will need to use the entire training set for this. In the code cell below, you will need to implement the following:\n",
    "- Import [`sklearn.grid_search.GridSearchCV`](http://scikit-learn.org/0.17/modules/generated/sklearn.grid_search.GridSearchCV.html) and [`sklearn.metrics.make_scorer`](http://scikit-learn.org/stable/modules/generated/sklearn.metrics.make_scorer.html).\n",
    "- Initialize the classifier you've chosen and store it in `clf`.\n",
    " - Set a `random_state` if one is available to the same state you set before.\n",
    "- Create a dictionary of parameters you wish to tune for the chosen model.\n",
    " - Example: `parameters = {'parameter' : [list of values]}`.\n",
    " - **Note:** Avoid tuning the `max_features` parameter of your learner if that parameter is available!\n",
    "- Use `make_scorer` to create an `fbeta_score` scoring object (with $\\beta = 0.5$).\n",
    "- Perform grid search on the classifier `clf` using the `'scorer'`, and store it in `grid_obj`.\n",
    "- Fit the grid search object to the training data (`X_train`, `y_train`), and store it in `grid_fit`.\n",
    "\n",
    "**Note:** Depending on the algorithm chosen and the parameter list, the following implementation may take some time to run!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unoptimized model\n",
      "------\n",
      "Accuracy score on testing data: 0.8297\n",
      "F-score on testing data: 0.6493\n",
      "\n",
      "Optimized Model\n",
      "------\n",
      "Final accuracy score on the testing data: 0.8415\n",
      "Final F-score on the testing data: 0.6861\n"
     ]
    }
   ],
   "source": [
    "# TODO: Import 'GridSearchCV', 'make_scorer', and any other necessary libraries\n",
    "from sklearn import grid_search\n",
    "from sklearn.metrics import make_scorer\n",
    "\n",
    "\n",
    "# TODO: Initialize the classifier\n",
    "clf = SGDClassifier(random_state = 7)\n",
    "\n",
    "\n",
    "# TODO: Create the parameters list you wish to tune, using a dictionary if needed.\n",
    "# HINT: parameters = {'parameter_1': [value1, value2], 'parameter_2': [value1, value2]}\n",
    "parameters = {'loss': ['hinge', 'log'], 'penalty': ['none', 'l2', 'l1', 'elasticnet'], 'alpha': [0.0001,0.001,0.01, 0.1, 1.0] }\n",
    "\n",
    "\n",
    "# TODO: Make an fbeta_score scoring object using make_scorer()\n",
    "scorer = make_scorer(fbeta_score, beta = 0.5)\n",
    "\n",
    "# TODO: Perform grid search on the classifier using 'scorer' as the scoring method using GridSearchCV()\n",
    "grid_obj = grid_search.GridSearchCV(estimator = clf, param_grid = parameters,scoring = scorer )\n",
    "\n",
    "# TODO: Fit the grid search object to the training data and find the optimal parameters using fit()\n",
    "grid_fit = grid_obj.fit(X_train, y_train)\n",
    "\n",
    "# Get the estimator\n",
    "best_clf = grid_fit.best_estimator_\n",
    "\n",
    "# Make predictions using the unoptimized and model\n",
    "predictions = (clf.fit(X_train, y_train)).predict(X_test)\n",
    "best_predictions = best_clf.predict(X_test)\n",
    "\n",
    "# Report the before-and-afterscores\n",
    "print(\"Unoptimized model\\n------\")\n",
    "print(\"Accuracy score on testing data: {:.4f}\".format(accuracy_score(y_test, predictions)))\n",
    "print(\"F-score on testing data: {:.4f}\".format(fbeta_score(y_test, predictions, beta = 0.5)))\n",
    "print(\"\\nOptimized Model\\n------\")\n",
    "print(\"Final accuracy score on the testing data: {:.4f}\".format(accuracy_score(y_test, best_predictions)))\n",
    "print(\"Final F-score on the testing data: {:.4f}\".format(fbeta_score(y_test, best_predictions, beta = 0.5)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 5 - Final Model Evaluation\n",
    "\n",
    "* What is your optimized model's accuracy and F-score on the testing data? \n",
    "* Are these scores better or worse than the unoptimized model? \n",
    "* How do the results from your optimized model compare to the naive predictor benchmarks you found earlier in **Question 1**?_  \n",
    "\n",
    "**Note:** Fill in the table below with your results, and then provide discussion in the **Answer** box."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Results:\n",
    "\n",
    "|     Metric     | Benchmark Predictor | Unoptimized Model | Optimized Model |\n",
    "| :------------: | :-----------------: | :---------------: | :-------------: | \n",
    "| Accuracy Score |    0.2478           |     0.8297        | 0.8415          |\n",
    "| F-score        |    0.2917           |   0.6493          |   0.6861        |\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer: **\n",
    "\n",
    "The optimised model delivered an accuracy score of 84.15%, which is 1.18% greater than the unoptimised models 82.97%.\n",
    "The optimised model delivered an F-score of 68.61%, which is 3.68% greater than the unoptimised models 64.93% score. \n",
    "Both these scores are better than the unoptimised models. \n",
    "\n",
    "The results from my optimised model are vastly superior to the benchmark predictor. It must be remembered, that this predictor is naive in that it always assumed an individual made more than $50,000. In such a scenario, there will be no true negatives or false negatives. This results in accuracy being the same as precision. \n",
    "It is encouraging to see that the optimised model is performing that much better than the benchmark, as it means the model and its parameters were the correct choice, and no errors with the data were made. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n",
    "## Feature Importance\n",
    "\n",
    "An important task when performing supervised learning on a dataset like the census data we study here is determining which features provide the most predictive power. By focusing on the relationship between only a few crucial features and the target label we simplify our understanding of the phenomenon, which is most always a useful thing to do. In the case of this project, that means we wish to identify a small number of features that most strongly predict whether an individual makes at most or more than \\$50,000.\n",
    "\n",
    "Choose a scikit-learn classifier (e.g., adaboost, random forests) that has a `feature_importance_` attribute, which is a function that ranks the importance of features according to the chosen classifier.  In the next python cell fit this classifier to training set and use this attribute to determine the top 5 most important features for the census dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 6 - Feature Relevance Observation\n",
    "When **Exploring the Data**, it was shown there are thirteen available features for each individual on record in the census data. Of these thirteen records, which five features do you believe to be most important for prediction, and in what order would you rank them and why?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer:**\n",
    "\n",
    "1. **education_level**: Education level could be a predictor for the level of skilled or unskilled work someone could obtain. For example, someone with a college degree in computer science could work for a software company, earning a large salary whereas someone who has only obtained a high school degree may work in a less specialised and qualified role, earning less. However, there may also be cases where people are self educated, just like a software engineer, so it is not always a guarantee. \n",
    "\n",
    "2. **occupation**: Following up with education, the actual occupation will also be useful to determine salary - for example, did you graduate college with an arts degree and go on to work in mid-level employment, or did you graduate as a doctor, who works as a head surgeon in a hospital?\n",
    "\n",
    "3. **hours-per-week**: This is, for those either employed by the state or privately, simple mathematics. For the most part, those who work 40-50 hours a week should be recieving more than those working say, 10 hours a week. However, occupation is an important factor in this, as you may be working 50 hours a week at McDonalds, or 10 hours a week as an investment banker. This is why I have rated occupation higher. \n",
    "\n",
    "4. **capital-gain**: What is their current monetary capital gains (indication of current wealth, and potential yearly earnings). I have ranked this above capital loss purely as there was far more populated data samples for capital-gain, and therefore more useful for modelling.\n",
    "\n",
    "5. **age**: Lastly, I have included age, as I feel experience and an opportunity to climb the ladder of progression should show some correlation with a persons age. For example someone who is out of University 5 years, as oppose to someone who has been out of University 20 years should show a trend in real earnings comparitively. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementation - Extracting Feature Importance\n",
    "Choose a `scikit-learn` supervised learning algorithm that has a `feature_importance_` attribute availble for it. This attribute is a function that ranks the importance of each feature when making predictions based on the chosen algorithm.\n",
    "\n",
    "In the code cell below, you will need to implement the following:\n",
    " - Import a supervised learning model from sklearn if it is different from the three used earlier.\n",
    " - Train the supervised model on the entire training set.\n",
    " - Extract the feature importances using `'.feature_importances_'`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAn4AAAFgCAYAAAAl5HQWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x114bfb350>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# TODO: Import a supervised learning model that has 'feature_importances_'\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "# TODO: Train the supervised model on the training set using .fit(X_train, y_train)\n",
    "model = RandomForestClassifier(random_state=7).fit(X_train, y_train)\n",
    "\n",
    "# TODO: Extract the feature importances using .feature_importances_ \n",
    "importances = model.feature_importances_\n",
    "\n",
    "# Plot\n",
    "vs.feature_plot(importances, X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 7 - Extracting Feature Importance\n",
    "\n",
    "Observe the visualization created above which displays the five most relevant features for predicting if an individual makes at most or above \\$50,000.  \n",
    "* How do these five features compare to the five features you discussed in **Question 6**?\n",
    "* If you were close to the same answer, how does this visualization confirm your thoughts? \n",
    "* If you were not close, why do you think these features are more relevant?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer:**\n",
    "\n",
    "Of the five features listed above (age, hours-per-week, capital-gain, marital-status, relationship), three of these were the same as mine (age, hour-per-week, capital-gain). This correlates with my understanding that with age, comes experience, and likely a large salary, that the hours-per-week you work is important (as most have contracted hours that reflects their individual salary) and that capital gain is a good indicator of percieved wealth.\n",
    "\n",
    "However, I was surprised to see both marital-status and relationship influence earnings so heavily. These features might be more relevant as it may reflect combined earnings, or that those who are succesful and earn more have the opportunity and money to be in a relationship.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Selection\n",
    "How does a model perform if we only use a subset of all the available features in the data? With less features required to train, the expectation is that training and prediction time is much lower — at the cost of performance metrics. From the visualization above, we see that the top five most important features contribute more than half of the importance of **all** features present in the data. This hints that we can attempt to *reduce the feature space* and simplify the information required for the model to learn. The code cell below will use the same optimized model you found earlier, and train it on the same training set *with only the top five important features*. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Final Model trained on full data\n",
      "------\n",
      "Accuracy on testing data: 0.8415\n",
      "F-score on testing data: 0.6861\n",
      "\n",
      "Final Model trained on reduced data\n",
      "------\n",
      "Accuracy on testing data: 0.7823\n",
      "F-score on testing data: 0.4826\n"
     ]
    }
   ],
   "source": [
    "# Import functionality for cloning a model\n",
    "from sklearn.base import clone\n",
    "\n",
    "# Reduce the feature space\n",
    "X_train_reduced = X_train[X_train.columns.values[(np.argsort(importances)[::-1])[:5]]]\n",
    "X_test_reduced = X_test[X_test.columns.values[(np.argsort(importances)[::-1])[:5]]]\n",
    "\n",
    "# Train on the \"best\" model found from grid search earlier\n",
    "clf = (clone(best_clf)).fit(X_train_reduced, y_train)\n",
    "\n",
    "# Make new predictions\n",
    "reduced_predictions = clf.predict(X_test_reduced)\n",
    "\n",
    "# Report scores from the final model using both versions of data\n",
    "print \"Final Model trained on full data\\n------\"\n",
    "print \"Accuracy on testing data: {:.4f}\".format(accuracy_score(y_test, best_predictions))\n",
    "print \"F-score on testing data: {:.4f}\".format(fbeta_score(y_test, best_predictions, beta = 0.5))\n",
    "print \"\\nFinal Model trained on reduced data\\n------\"\n",
    "print \"Accuracy on testing data: {:.4f}\".format(accuracy_score(y_test, reduced_predictions))\n",
    "print \"F-score on testing data: {:.4f}\".format(fbeta_score(y_test, reduced_predictions, beta = 0.5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 8 - Effects of Feature Selection\n",
    "\n",
    "* How does the final model's F-score and accuracy score on the reduced data using only five features compare to those same scores when all features are used?\n",
    "* If training time was a factor, would you consider using the reduced data as your training set?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer:**\n",
    "\n",
    "The accuracy drop using the reduced feature set is good, seeing around a 6% drop in accuracy when compared with the full 13 feature set. I would be happy to use this lowered accuracy score if training time was a factor, as 78% is still a great result for removing over half the features (5 Vs 13).\n",
    "\n",
    "However, the F-score sees a substantially greater drop, falling by just over 20%. This is not acceptable for me, as we want an f-score as close to possible as 1. Previously, with the full 13 features and optimisation, we were achieving close to 70% on our f-score. With the reduced training set, we are not even getting 50%. One option would be to consider other models to see if they would provide a better tuned score with the reduced set that I previously listed as too slow (SVM) on the full feature set. If SVC could provide an f-score close to the 68% figure, with reduced training time, then I would consider using the reduced training set.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> **Note**: Once you have completed all of the code implementations and successfully answered each question above, you may finalize your work by exporting the iPython Notebook as an HTML document. You can do this by using the menu above and navigating to  \n",
    "**File -> Download as -> HTML (.html)**. Include the finished document along with this notebook as your submission."
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Raw Cell Format",
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}