Module-01c-Model-Regression.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "# Transform and Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Let us build a Regression Model for prediciting the amount to be approved"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Load the libraries\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Default Variables\n",
    "%matplotlib inline\n",
    "plt.rcParams['figure.figsize'] = (16,9)\n",
    "plt.rcParams['font.size'] = 18\n",
    "plt.style.use('fivethirtyeight')\n",
    "pd.set_option('display.float_format', lambda x: '%.2f' % x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Load the dataset\n",
    "df = pd.read_csv(\"data/loan_data_clean.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>default</th>\n",
       "      <th>amount</th>\n",
       "      <th>interest</th>\n",
       "      <th>grade</th>\n",
       "      <th>years</th>\n",
       "      <th>ownership</th>\n",
       "      <th>income</th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>5000</td>\n",
       "      <td>10.65</td>\n",
       "      <td>B</td>\n",
       "      <td>10.00</td>\n",
       "      <td>RENT</td>\n",
       "      <td>24000.00</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>2400</td>\n",
       "      <td>10.99</td>\n",
       "      <td>C</td>\n",
       "      <td>25.00</td>\n",
       "      <td>RENT</td>\n",
       "      <td>12252.00</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>10000</td>\n",
       "      <td>13.49</td>\n",
       "      <td>C</td>\n",
       "      <td>13.00</td>\n",
       "      <td>RENT</td>\n",
       "      <td>49200.00</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>5000</td>\n",
       "      <td>10.99</td>\n",
       "      <td>A</td>\n",
       "      <td>3.00</td>\n",
       "      <td>RENT</td>\n",
       "      <td>36000.00</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>3000</td>\n",
       "      <td>10.99</td>\n",
       "      <td>E</td>\n",
       "      <td>9.00</td>\n",
       "      <td>RENT</td>\n",
       "      <td>48000.00</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   default  amount  interest grade  years ownership   income  age\n",
       "0        0    5000     10.65     B  10.00      RENT 24000.00   33\n",
       "1        0    2400     10.99     C  25.00      RENT 12252.00   31\n",
       "2        0   10000     13.49     C  13.00      RENT 49200.00   24\n",
       "3        0    5000     10.99     A   3.00      RENT 36000.00   39\n",
       "4        0    3000     10.99     E   9.00      RENT 48000.00   24"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Transform Variables\n",
    "\n",
    "Let us create feature and target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Select the initial feature set\n",
    "X_raw = df[['age', 'grade', 'years', 'ownership', 'income']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Convert the categorical variables in to numerical values\n",
    "from sklearn.preprocessing import OneHotEncoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Create the feature set X\n",
    "X = pd.get_dummies(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Create the target from amount and default\n",
    "y = df.amount * (1 - df.default)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0     5000\n",
       "1     2400\n",
       "2    10000\n",
       "3     5000\n",
       "4     3000\n",
       "dtype: int64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Build Model - Linear Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# import the sklearn linear model\n",
    "from sklearn.linear_model import LinearRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# initiate the Linear Regression Model\n",
    "model_ols = LinearRegression(normalize = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=True)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Review the parameters in the Linear Regression\n",
    "model_ols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=True)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_ols.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  8.22428831e+00,   3.56559740e+01,   3.41152469e-02,\n",
       "        -5.56093383e+02,   4.63115647e+02,  -5.27037725e+02,\n",
       "         5.13726114e+02,   1.53946293e+03,   9.03022126e+02,\n",
       "         2.15733968e+03,   4.37125155e+02,  -6.26484704e+02,\n",
       "        -2.02390459e+02,  -3.56487860e+02])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# What are the coeffecients of the model\n",
    "model_ols.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5867.0568590363873"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_ols.intercept_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Calculate Model - Predictions & Error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# predict the y\n",
    "y_pred = model_ols.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "40135325.120906696"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# import metrics from sklearn\n",
    "np.sum((y_pred - y)**2)/y.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import mean_squared_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "40135325.120906696"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculate mean squared error\n",
    "mean_squared_error(y_pred,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6335.2446772722751"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Root mean square error\n",
    "np.sqrt(np.sum((y_pred - y)**2)/y.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBsAAAIrCAYAAACu6JFTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqGAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x119fd22b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(y, bins=50, alpha=0.4, label=\"true\")\n",
    "plt.hist(y_pred, bins=50, alpha=0.4, label=\"pred\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBsAAAIrCAYAAACu6JFTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDQAAAAAAwFEEGwAAAAAAgKP+P44b78nvrtTTAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x119fd8780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(y_pred, y)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Evaluate Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.098018899623945832"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# What is the score given by the model\n",
    "model_ols.score(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6335.2446772722751"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# What is the root mean square error\n",
    "np.sqrt(mean_squared_error(y_pred, y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6670.7111933826272"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How does rmse compare with standard deviation of the target\n",
    "y.std()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Generalisation Error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Get the module for train test split\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "train_test_split?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Split the data in test and training - 20% and 80%\n",
    "X_train, X_test, y_train, y_test = train_test_split(X,y, test_size = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((23272, 14), (5819, 14), (23272,), (5819,))"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Initiate the model\n",
    "model_general = LinearRegression(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=True)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Fit the model\n",
    "model_general.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Make predictions for test and train\n",
    "y_pred_general_train = model_general.predict(X_train)\n",
    "y_pred_general_test = model_general.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Find the errors for test and train\n",
    "mse_general_train = mean_squared_error(y_pred_general_train, y_train)\n",
    "mse_general_test = mean_squared_error(y_pred_general_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(769603.69098217785, 40303216.575326569, 39533612.884344392)"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Find the generalisation error\n",
    "mse_general_train - mse_general_test, mse_general_train, mse_general_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Build Complex Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Import Polynominal Features\n",
    "from sklearn.preprocessing import PolynomialFeatures"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Initiate Polynominal Features for Degree 2\n",
    "poly = PolynomialFeatures(degree=2, interaction_only=True, include_bias=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Create Polynominal Features\n",
    "X_poly = poly.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((29091, 14), (29091, 105))"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# See the new dataset\n",
    "X.shape, X_poly.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Create split and train\n",
    "X_train_poly, X_test_poly, y_train_poly, y_test_poly = train_test_split(X_poly,y, test_size = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Initiate the model\n",
    "model_poly = LinearRegression(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=True)"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fit the model\n",
    "model_poly.fit(X_train_poly, y_train_poly)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Make predictions for test and train\n",
    "y_pred_poly_train = model_poly.predict(X_train_poly) \n",
    "y_pred_poly_test = model_poly.predict(X_test_poly) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Find the errors for test and train\n",
    "mse_poly_train = mean_squared_error(y_pred_poly_train,y_train_poly )\n",
    "mse_poly_test = mean_squared_error(y_pred_poly_test,y_test_poly )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2152944.5680075735, 38314644.263922311, 40467588.831929885)"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Find the generalisation error\n",
    "mse_poly_test - mse_poly_train, mse_poly_train, mse_poly_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(769603.69098217785, 40303216.575326569, 39533612.884344392)"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mse_general_train - mse_general_test, mse_general_train, mse_general_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "For Discussion\n",
    "- Why has the generalisation error gone up?\n",
    "- Should a complex model perform better than a simple one? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Regularization - Ridge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Get ridge regression from linear_models\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Initiate model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Fit the model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Make predictions for test and train\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "#Find the errors for test and train\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Find the generalisation error\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Cross Validation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "Finding alpha using Cross Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Get ridge regression from linear_models\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Initiate model with alphas = 0.1, 0.001, 0.0001\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Fit the model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Find the correct alpha\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "## Exercise: Regularization - Lasso"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}