5-EX-Regression-K-fold-Cross-Validation-v2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creative Commons CC BY 4.0 Lynd Bacon & Associates, Ltd. Not warranted to be suitable for any particular purpose. (You're on your own!)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# K-Fold Cross-Validation for Regression\n",
    "\n",
    "In the following we're going to apply the _K-fold cross-validation_ method to assess model performance on data that the model \"hasn't seen.\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preliminaries\n",
    "\n",
    "Let's first do some set-ups, and get some packages we're going to need."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "InteractiveShell.ast_node_interactivity = \"all\"\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn import linear_model  # OLS \n",
    "from sklearn.metrics import mean_squared_error, r2_score, SCORERS # Basic metrics\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import cross_validate, cross_val_predict"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Input and Preprocessing\n",
    "\n",
    "Let's get the patient satisfaction data in the file `DECART-patSat.csv` into a Pandas DataFrame, and then extract into a numpy array the data we want to use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['caseID', 'patSat', 'q2', 'q3', 'q4', 'q5', 'q6', 'q7', 'q8', 'q9',\n",
       "       'ptCat'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ptSatDF=pd.read_csv('DATA/ML/DECART-patSat.csv')\n",
    "ptSatDF.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "That `ptCat` variable is categorical:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    679\n",
       "0    596\n",
       "2    536\n",
       "Name: ptCat, dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ptSatDF.ptCat.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dummy Coding, or \"OneHot\" Coding\n",
    "\n",
    "To use `ptCat` in a regression model we need to code it somehow.  Let's \"dummy\" (zero/one) code it, omitting the first category:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['patSat', 'q2', 'q3', 'q4', 'q5', 'q6', 'q7', 'q8', 'q9', 'ptCat1',\n",
       "       'ptCat2'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ptSatDF[['ptCat1','ptCat2']]=pd.get_dummies(ptSatDF.ptCat,drop_first=True)\n",
    "ptSatDF2=ptSatDF.drop(['ptCat','caseID'],axis=1)\n",
    "ptSatDF2.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Effects Coding of Categorical Predictor Variables Digression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we wanted _effects_ coding of `ptCat` instead of dummy coding, we could just dummy code first, omitting the reference category, and then replace every pair of zeros with a pair of -1's.   \n",
    "\n",
    "See if the following works to do this for you, or if it produces unexpected results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>255</td>\n",
       "      <td>255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     1    2\n",
       "0  255  255\n",
       "1    0    1\n",
       "3    1    0"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ptCat=ptSatDF.ptCat.to_numpy()   # select the ptCat col\n",
    "pcEC=pd.get_dummies(ptCat,drop_first=True)\n",
    "rowTot=pcEC.apply(sum,axis=1)  # calc row totals\n",
    "a=-1\n",
    "pcEC.loc[rowTot==0]=a\n",
    "pcEC.drop_duplicates()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating numpy Arrays for CV Training and Testing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's now make the numpy objects we're going to feed into our regression model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1811, 10)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "(1811,)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ptSatX=ptSatDF2.iloc[:,1:].to_numpy()  # predictor vars\n",
    "ptSatY=ptSatDF2.iloc[:,0].to_numpy()   # pt satisfaction dep var\n",
    "ptSatX.shape  # Just checking\n",
    "ptSatY.shape\n",
    "ptSatY.min()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " # A Simple Way To Do K Fold Cross-Validation\n",
    "\n",
    "Here we do 10 fold CV, retaining fit and timing metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(57)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Available Metrics\n",
    "\n",
    "These are for a variety of \"learner\" model types."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Here's what the available metrics are. Uncomment and exec to list them.\n",
    "# sorted(SCORERS.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "regr=linear_model.LinearRegression()  # lin reg model w/ defaults\n",
    "scoring=['r2','neg_mean_squared_error','explained_variance'] # list of metrics to be returned\n",
    "scores = cross_validate(regr, ptSatX, ptSatY, cv=10,\n",
    "                               scoring=scoring,\n",
    "                       return_train_score=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`scores` should be a python dict.  We can make it easier to inspect by making a DataFrame using it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fit_time</th>\n",
       "      <th>score_time</th>\n",
       "      <th>test_r2</th>\n",
       "      <th>train_r2</th>\n",
       "      <th>test_neg_mean_squared_error</th>\n",
       "      <th>train_neg_mean_squared_error</th>\n",
       "      <th>test_explained_variance</th>\n",
       "      <th>train_explained_variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.001269</td>\n",
       "      <td>0.001168</td>\n",
       "      <td>0.744029</td>\n",
       "      <td>0.701565</td>\n",
       "      <td>-1.811215</td>\n",
       "      <td>-1.919569</td>\n",
       "      <td>0.744030</td>\n",
       "      <td>0.701565</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.000879</td>\n",
       "      <td>0.000907</td>\n",
       "      <td>0.692256</td>\n",
       "      <td>0.707983</td>\n",
       "      <td>-2.206559</td>\n",
       "      <td>-1.876766</td>\n",
       "      <td>0.692308</td>\n",
       "      <td>0.707983</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000824</td>\n",
       "      <td>0.000896</td>\n",
       "      <td>0.731350</td>\n",
       "      <td>0.704014</td>\n",
       "      <td>-1.600421</td>\n",
       "      <td>-1.941993</td>\n",
       "      <td>0.731660</td>\n",
       "      <td>0.704014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.000833</td>\n",
       "      <td>0.000896</td>\n",
       "      <td>0.643689</td>\n",
       "      <td>0.712851</td>\n",
       "      <td>-2.194789</td>\n",
       "      <td>-1.877032</td>\n",
       "      <td>0.649401</td>\n",
       "      <td>0.712851</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.001951</td>\n",
       "      <td>0.000885</td>\n",
       "      <td>0.683885</td>\n",
       "      <td>0.708667</td>\n",
       "      <td>-2.001926</td>\n",
       "      <td>-1.898538</td>\n",
       "      <td>0.684635</td>\n",
       "      <td>0.708667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.001951</td>\n",
       "      <td>0.000891</td>\n",
       "      <td>0.668084</td>\n",
       "      <td>0.710313</td>\n",
       "      <td>-2.085132</td>\n",
       "      <td>-1.888684</td>\n",
       "      <td>0.672082</td>\n",
       "      <td>0.710313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.001938</td>\n",
       "      <td>0.000890</td>\n",
       "      <td>0.715595</td>\n",
       "      <td>0.705548</td>\n",
       "      <td>-1.710859</td>\n",
       "      <td>-1.928804</td>\n",
       "      <td>0.716347</td>\n",
       "      <td>0.705548</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.001943</td>\n",
       "      <td>0.000894</td>\n",
       "      <td>0.687332</td>\n",
       "      <td>0.708571</td>\n",
       "      <td>-2.324607</td>\n",
       "      <td>-1.862466</td>\n",
       "      <td>0.687441</td>\n",
       "      <td>0.708571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.001950</td>\n",
       "      <td>0.000894</td>\n",
       "      <td>0.680903</td>\n",
       "      <td>0.708593</td>\n",
       "      <td>-1.912441</td>\n",
       "      <td>-1.907681</td>\n",
       "      <td>0.680906</td>\n",
       "      <td>0.708593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.001962</td>\n",
       "      <td>0.000899</td>\n",
       "      <td>0.742841</td>\n",
       "      <td>0.702754</td>\n",
       "      <td>-1.600901</td>\n",
       "      <td>-1.940552</td>\n",
       "      <td>0.743057</td>\n",
       "      <td>0.702754</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   fit_time  score_time   test_r2  train_r2  test_neg_mean_squared_error  \\\n",
       "0  0.001269    0.001168  0.744029  0.701565                    -1.811215   \n",
       "1  0.000879    0.000907  0.692256  0.707983                    -2.206559   \n",
       "2  0.000824    0.000896  0.731350  0.704014                    -1.600421   \n",
       "3  0.000833    0.000896  0.643689  0.712851                    -2.194789   \n",
       "4  0.001951    0.000885  0.683885  0.708667                    -2.001926   \n",
       "5  0.001951    0.000891  0.668084  0.710313                    -2.085132   \n",
       "6  0.001938    0.000890  0.715595  0.705548                    -1.710859   \n",
       "7  0.001943    0.000894  0.687332  0.708571                    -2.324607   \n",
       "8  0.001950    0.000894  0.680903  0.708593                    -1.912441   \n",
       "9  0.001962    0.000899  0.742841  0.702754                    -1.600901   \n",
       "\n",
       "   train_neg_mean_squared_error  test_explained_variance  \\\n",
       "0                     -1.919569                 0.744030   \n",
       "1                     -1.876766                 0.692308   \n",
       "2                     -1.941993                 0.731660   \n",
       "3                     -1.877032                 0.649401   \n",
       "4                     -1.898538                 0.684635   \n",
       "5                     -1.888684                 0.672082   \n",
       "6                     -1.928804                 0.716347   \n",
       "7                     -1.862466                 0.687441   \n",
       "8                     -1.907681                 0.680906   \n",
       "9                     -1.940552                 0.743057   \n",
       "\n",
       "   train_explained_variance  \n",
       "0                  0.701565  \n",
       "1                  0.707983  \n",
       "2                  0.704014  \n",
       "3                  0.712851  \n",
       "4                  0.708667  \n",
       "5                  0.710313  \n",
       "6                  0.705548  \n",
       "7                  0.708571  \n",
       "8                  0.708593  \n",
       "9                  0.702754  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# type(scores)\n",
    "# scores.keys()\n",
    "# scores.values()\n",
    "pd.DataFrame(scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Summarizing the folds' statistics:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fit_time</th>\n",
       "      <th>score_time</th>\n",
       "      <th>test_r2</th>\n",
       "      <th>train_r2</th>\n",
       "      <th>test_neg_mean_squared_error</th>\n",
       "      <th>train_neg_mean_squared_error</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>10.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>10.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.001304</td>\n",
       "      <td>0.000604</td>\n",
       "      <td>0.698996</td>\n",
       "      <td>0.707086</td>\n",
       "      <td>-1.944885</td>\n",
       "      <td>-1.904209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.000466</td>\n",
       "      <td>0.000144</td>\n",
       "      <td>0.033374</td>\n",
       "      <td>0.003535</td>\n",
       "      <td>0.260065</td>\n",
       "      <td>0.028104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000618</td>\n",
       "      <td>0.000323</td>\n",
       "      <td>0.643689</td>\n",
       "      <td>0.701565</td>\n",
       "      <td>-2.324607</td>\n",
       "      <td>-1.941993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.001112</td>\n",
       "      <td>0.000584</td>\n",
       "      <td>0.681648</td>\n",
       "      <td>0.704397</td>\n",
       "      <td>-2.167375</td>\n",
       "      <td>-1.926495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.001200</td>\n",
       "      <td>0.000628</td>\n",
       "      <td>0.689794</td>\n",
       "      <td>0.708277</td>\n",
       "      <td>-1.957183</td>\n",
       "      <td>-1.903110</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.001415</td>\n",
       "      <td>0.000649</td>\n",
       "      <td>0.727411</td>\n",
       "      <td>0.708648</td>\n",
       "      <td>-1.735948</td>\n",
       "      <td>-1.879945</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.002412</td>\n",
       "      <td>0.000875</td>\n",
       "      <td>0.744029</td>\n",
       "      <td>0.712851</td>\n",
       "      <td>-1.600421</td>\n",
       "      <td>-1.862466</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        fit_time  score_time    test_r2   train_r2  \\\n",
       "count  10.000000   10.000000  10.000000  10.000000   \n",
       "mean    0.001304    0.000604   0.698996   0.707086   \n",
       "std     0.000466    0.000144   0.033374   0.003535   \n",
       "min     0.000618    0.000323   0.643689   0.701565   \n",
       "25%     0.001112    0.000584   0.681648   0.704397   \n",
       "50%     0.001200    0.000628   0.689794   0.708277   \n",
       "75%     0.001415    0.000649   0.727411   0.708648   \n",
       "max     0.002412    0.000875   0.744029   0.712851   \n",
       "\n",
       "       test_neg_mean_squared_error  train_neg_mean_squared_error  \n",
       "count                    10.000000                     10.000000  \n",
       "mean                     -1.944885                     -1.904209  \n",
       "std                       0.260065                      0.028104  \n",
       "min                      -2.324607                     -1.941993  \n",
       "25%                      -2.167375                     -1.926495  \n",
       "50%                      -1.957183                     -1.903110  \n",
       "75%                      -1.735948                     -1.879945  \n",
       "max                      -1.600421                     -1.862466  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(scores).describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predictions From Cross-Validated Model\n",
    "\n",
    "Predicted values for the dependent measure based on the _test_ cv folds can be obtained for Dx purposes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predicted patient satisfaction ratings\n",
    "\n",
    "# regrCVpreds will be a 1D array of predicted values based on test data\n",
    " \n",
    "regrCVPreds=cross_val_predict(regr,ptSatX,ptSatY,cv=10,n_jobs=-1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot of Predicted Based on Test Folds vs. Actual\n",
    "\n",
    "Note the semicolon at the end of the last line.  Remove it and run the following cell again to see what it does."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "# points=plt.scatter(ptSatY,regrCVPreds,alpha=0.1)\n",
    "plt.figure(figsize=(8,8))\n",
    "plot=sns.regplot(ptSatY,regrCVPreds,\n",
    "                 scatter=True,\n",
    "                 color='.1',x_jitter=.2,\n",
    "                 scatter_kws={\"alpha\":0.4,\"color\":\"grey\"})\n",
    "\n",
    "plt.xlim(0.5,10.5)\n",
    "\n",
    "plot=plot.set(xlabel='PT Sat Rating',ylabel='Predicted Rating');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Another Way to Do the Kfold Shuffle\n",
    "\n",
    "Let's use the `radon` data and try cross-validation in a slightly different way."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "radon=pd.read_csv('DATA/radon.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Munging\n",
    "\n",
    "We've seen that there is a row of data that's missing a value for the `hhincome` variable.  Let's kick that observation to the curb, and work with the rest of the data.\n",
    "\n",
    "The following stmt selects all rows that aren't missing a value for `hhincome`, gets columns `lcanmort` through `hhincome`, and drops columns `radon`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2880, 7)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Index(['lcanmort', 'lnradon', 'obesity', 'over65', 'cursmoke', 'evrsmoke',\n",
       "       'hhincome'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "radon2=radon.loc[radon.hhincome.notnull(),'lcanmort':'hhincome'] \\\n",
    "        .drop('radon',axis=1)\n",
    "radon2.shape\n",
    "radon2.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 20 Fold CV\n",
    "\n",
    "Next will set up to do 20 fold cross-validation, pre-shuffling the data, and setting the random seed to a constant so that results can be replicated if need be."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "kf=KFold(n_splits=20,random_state=99,shuffle=True)\n",
    "X=radon2.iloc[:,1:].to_numpy()\n",
    "y=radon2.loc[:,'lcanmort'].to_numpy()\n",
    "\n",
    "cvres=list()  # This will hold cv results\n",
    "\n",
    "regr=linear_model.LinearRegression() # define a reg model to use\n",
    "\n",
    "for traindx, testdx in kf.split(X):  # loop over folds\n",
    "    resDict={}                       # Dictionary to hold fold results\n",
    "    XTrain=X[traindx]                # Xtrain\n",
    "    yTrain=y[traindx]\n",
    "    XTest=X[testdx]\n",
    "    yTest=y[testdx]\n",
    "    regModel=regr.fit(XTrain,yTrain)\n",
    "    trainPred=regModel.predict(XTrain)\n",
    "    trainR2=r2_score(yTrain,trainPred)\n",
    "    trainMSE=mean_squared_error(yTrain,trainPred)\n",
    "    testPred=regModel.predict(XTest)\n",
    "    testR2=r2_score(yTest,testPred)\n",
    "    testMSE=mean_squared_error(yTest,testPred)\n",
    "    resDict.update({'trainR2':trainR2,\n",
    "                    'testR2':testR2,\n",
    "                    'trainMSE':trainMSE,\n",
    "                    'testMSE':testMSE})\n",
    "    cvres.append(resDict)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Rearranging cols to make train vs test comparisons easier\n",
    "\n",
    "cvresDF=pd.DataFrame(cvres)[['trainMSE','testMSE','trainR2','testR2']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inspecting Results\n",
    "\n",
    "Let's summarize results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>trainMSE</th>\n",
       "      <th>testMSE</th>\n",
       "      <th>trainR2</th>\n",
       "      <th>testR2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>20.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>20.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>166.338463</td>\n",
       "      <td>167.178183</td>\n",
       "      <td>0.465881</td>\n",
       "      <td>0.458306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.504895</td>\n",
       "      <td>28.563857</td>\n",
       "      <td>0.003490</td>\n",
       "      <td>0.063702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>162.079522</td>\n",
       "      <td>126.813798</td>\n",
       "      <td>0.459364</td>\n",
       "      <td>0.333667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>165.744857</td>\n",
       "      <td>149.900947</td>\n",
       "      <td>0.464056</td>\n",
       "      <td>0.430661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>166.595322</td>\n",
       "      <td>162.365318</td>\n",
       "      <td>0.465750</td>\n",
       "      <td>0.465242</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>167.248859</td>\n",
       "      <td>178.546491</td>\n",
       "      <td>0.467429</td>\n",
       "      <td>0.495068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>168.463209</td>\n",
       "      <td>247.957006</td>\n",
       "      <td>0.472399</td>\n",
       "      <td>0.577696</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         trainMSE     testMSE    trainR2     testR2\n",
       "count   20.000000   20.000000  20.000000  20.000000\n",
       "mean   166.338463  167.178183   0.465881   0.458306\n",
       "std      1.504895   28.563857   0.003490   0.063702\n",
       "min    162.079522  126.813798   0.459364   0.333667\n",
       "25%    165.744857  149.900947   0.464056   0.430661\n",
       "50%    166.595322  162.365318   0.465750   0.465242\n",
       "75%    167.248859  178.546491   0.467429   0.495068\n",
       "max    168.463209  247.957006   0.472399   0.577696"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cvresDF.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compare train and test metrics directly by looking at the differences betwen the within fold metrics."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "cvresDF['TestTrainMSEDiff']=cvresDF.testMSE-cvresDF.trainMSE  # This is test-train\n",
    "cvresDF['TrainTestR2Diff']=cvresDF.trainR2-cvresDF.testR2     # This is train-test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TestTrainMSEDiff</th>\n",
       "      <th>TrainTestR2Diff</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>20.000000</td>\n",
       "      <td>20.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.839720</td>\n",
       "      <td>0.007575</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>30.068707</td>\n",
       "      <td>0.067153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-41.649411</td>\n",
       "      <td>-0.118332</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-17.347912</td>\n",
       "      <td>-0.031012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>-4.230004</td>\n",
       "      <td>0.000309</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>12.801634</td>\n",
       "      <td>0.036768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>85.877483</td>\n",
       "      <td>0.137989</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       TestTrainMSEDiff  TrainTestR2Diff\n",
       "count         20.000000        20.000000\n",
       "mean           0.839720         0.007575\n",
       "std           30.068707         0.067153\n",
       "min          -41.649411        -0.118332\n",
       "25%          -17.347912        -0.031012\n",
       "50%           -4.230004         0.000309\n",
       "75%           12.801634         0.036768\n",
       "max           85.877483         0.137989"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cvresDF[['TestTrainMSEDiff','TrainTestR2Diff']].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# UDU:  Repeat the K Fold Procedure, Above, Using the pt Sat Data"
   ]
  }
 ],
 "metadata": {
  "hide_input": false,
  "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": "3.6.8"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "290px"
   },
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}