Main_babyweight.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 819,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.linear_model import LinearRegression\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 1: Prepare the Workspace"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 820,
   "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>ID</th>\n",
       "      <th>SEX</th>\n",
       "      <th>MARITAL</th>\n",
       "      <th>FAGE</th>\n",
       "      <th>GAINED</th>\n",
       "      <th>VISITS</th>\n",
       "      <th>MAGE</th>\n",
       "      <th>FEDUC</th>\n",
       "      <th>MEDUC</th>\n",
       "      <th>TOTALP</th>\n",
       "      <th>...</th>\n",
       "      <th>HYPERCH</th>\n",
       "      <th>HYPERPR</th>\n",
       "      <th>ECLAMP</th>\n",
       "      <th>CERVIX</th>\n",
       "      <th>PINFANT</th>\n",
       "      <th>PRETERM</th>\n",
       "      <th>RENAL</th>\n",
       "      <th>RHSEN</th>\n",
       "      <th>UTERINE</th>\n",
       "      <th>BWEIGHT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "      <td>101400.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>52700.500000</td>\n",
       "      <td>1.485671</td>\n",
       "      <td>1.303817</td>\n",
       "      <td>30.174477</td>\n",
       "      <td>30.283037</td>\n",
       "      <td>12.436943</td>\n",
       "      <td>27.736312</td>\n",
       "      <td>12.926893</td>\n",
       "      <td>13.256489</td>\n",
       "      <td>2.378462</td>\n",
       "      <td>...</td>\n",
       "      <td>0.013767</td>\n",
       "      <td>0.050039</td>\n",
       "      <td>0.003807</td>\n",
       "      <td>0.003738</td>\n",
       "      <td>0.005680</td>\n",
       "      <td>0.008856</td>\n",
       "      <td>0.002189</td>\n",
       "      <td>0.003777</td>\n",
       "      <td>0.003442</td>\n",
       "      <td>7.258066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>29271.802985</td>\n",
       "      <td>0.500349</td>\n",
       "      <td>0.459907</td>\n",
       "      <td>6.775576</td>\n",
       "      <td>13.615401</td>\n",
       "      <td>3.728901</td>\n",
       "      <td>5.957369</td>\n",
       "      <td>2.926569</td>\n",
       "      <td>2.932693</td>\n",
       "      <td>1.490272</td>\n",
       "      <td>...</td>\n",
       "      <td>0.116524</td>\n",
       "      <td>0.218027</td>\n",
       "      <td>0.061581</td>\n",
       "      <td>0.061022</td>\n",
       "      <td>0.075155</td>\n",
       "      <td>0.093689</td>\n",
       "      <td>0.046739</td>\n",
       "      <td>0.061342</td>\n",
       "      <td>0.058566</td>\n",
       "      <td>1.329461</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>2001.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.187500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>27350.750000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>21.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>23.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>6.625000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>52700.500000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>28.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>7.375000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>78050.250000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>35.000000</td>\n",
       "      <td>39.000000</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>32.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>8.062500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>103400.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>74.000000</td>\n",
       "      <td>98.000000</td>\n",
       "      <td>49.000000</td>\n",
       "      <td>53.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>17.000000</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>13.062500</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 35 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  ID            SEX        MARITAL           FAGE  \\\n",
       "count  101400.000000  101400.000000  101400.000000  101400.000000   \n",
       "mean    52700.500000       1.485671       1.303817      30.174477   \n",
       "std     29271.802985       0.500349       0.459907       6.775576   \n",
       "min      2001.000000       1.000000       1.000000      14.000000   \n",
       "25%     27350.750000       1.000000       1.000000      25.000000   \n",
       "50%     52700.500000       1.000000       1.000000      30.000000   \n",
       "75%     78050.250000       2.000000       2.000000      35.000000   \n",
       "max    103400.000000       9.000000       2.000000      74.000000   \n",
       "\n",
       "              GAINED         VISITS           MAGE          FEDUC  \\\n",
       "count  101400.000000  101400.000000  101400.000000  101400.000000   \n",
       "mean       30.283037      12.436943      27.736312      12.926893   \n",
       "std        13.615401       3.728901       5.957369       2.926569   \n",
       "min         0.000000       0.000000      11.000000       0.000000   \n",
       "25%        21.000000      10.000000      23.000000      12.000000   \n",
       "50%        30.000000      12.000000      28.000000      12.000000   \n",
       "75%        39.000000      15.000000      32.000000      16.000000   \n",
       "max        98.000000      49.000000      53.000000      17.000000   \n",
       "\n",
       "               MEDUC         TOTALP  ...        HYPERCH        HYPERPR  \\\n",
       "count  101400.000000  101400.000000  ...  101400.000000  101400.000000   \n",
       "mean       13.256489       2.378462  ...       0.013767       0.050039   \n",
       "std         2.932693       1.490272  ...       0.116524       0.218027   \n",
       "min         0.000000       1.000000  ...       0.000000       0.000000   \n",
       "25%        12.000000       1.000000  ...       0.000000       0.000000   \n",
       "50%        13.000000       2.000000  ...       0.000000       0.000000   \n",
       "75%        16.000000       3.000000  ...       0.000000       0.000000   \n",
       "max        17.000000      20.000000  ...       1.000000       1.000000   \n",
       "\n",
       "              ECLAMP         CERVIX        PINFANT        PRETERM  \\\n",
       "count  101400.000000  101400.000000  101400.000000  101400.000000   \n",
       "mean        0.003807       0.003738       0.005680       0.008856   \n",
       "std         0.061581       0.061022       0.075155       0.093689   \n",
       "min         0.000000       0.000000       0.000000       0.000000   \n",
       "25%         0.000000       0.000000       0.000000       0.000000   \n",
       "50%         0.000000       0.000000       0.000000       0.000000   \n",
       "75%         0.000000       0.000000       0.000000       0.000000   \n",
       "max         1.000000       1.000000       1.000000       1.000000   \n",
       "\n",
       "               RENAL          RHSEN        UTERINE        BWEIGHT  \n",
       "count  101400.000000  101400.000000  101400.000000  101400.000000  \n",
       "mean        0.002189       0.003777       0.003442       7.258066  \n",
       "std         0.046739       0.061342       0.058566       1.329461  \n",
       "min         0.000000       0.000000       0.000000       0.187500  \n",
       "25%         0.000000       0.000000       0.000000       6.625000  \n",
       "50%         0.000000       0.000000       0.000000       7.375000  \n",
       "75%         0.000000       0.000000       0.000000       8.062500  \n",
       "max         1.000000       1.000000       1.000000      13.062500  \n",
       "\n",
       "[8 rows x 35 columns]"
      ]
     },
     "execution_count": 820,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Atrain1 = pd.read_csv(\"baby-weights-dataset2.csv\")\n",
    "#this gives information about babies born, together with birth weight\n",
    "Atrain1.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 2: Data Exploration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 821,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BWEIGHT Variable\n",
      "Mean: 7.2580658284023665\n",
      "stdev  1.3294541253578183\n",
      "Min: 0.1875\n",
      "Max: 13.0625\n",
      "25% percentile:  6.625\n",
      "Median: 7.375\n",
      "75% Percentile:  8.0625\n"
     ]
    }
   ],
   "source": [
    "print(\"BWEIGHT Variable\")\n",
    "print(\"Mean:\",Atrain1['BWEIGHT'].mean())\n",
    "print(\"stdev \",np.std(Atrain1['BWEIGHT']))\n",
    "print(\"Min:\", Atrain1['BWEIGHT'].min())\n",
    "print(\"Max:\",Atrain1['BWEIGHT'].max())\n",
    "print(\"25% percentile: \",np.percentile(Atrain1['BWEIGHT'], 25))\n",
    "print(\"Median:\",np.percentile(Atrain1['BWEIGHT'], 50))\n",
    "print(\"75% Percentile: \",np.percentile(Atrain1['BWEIGHT'], 75))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 822,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7ffe88e36150>"
      ]
     },
     "execution_count": 822,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot the histogram of the BWEIGHT variable\n",
    "sns.distplot(Atrain1['BWEIGHT'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 823,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: -0.936593\n",
      "Kurtosis: 2.868780\n"
     ]
    }
   ],
   "source": [
    "#Computing the Skewness and the Kurtosis of the Birth Weight variable\n",
    "print(\"Skewness: %f\" % Atrain1['BWEIGHT'].skew())\n",
    "print(\"Kurtosis: %f\" % Atrain1['BWEIGHT'].kurt())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 824,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "quantitative : ['SEX', 'MARITAL', 'FAGE', 'GAINED', 'VISITS', 'MAGE', 'FEDUC', 'MEDUC', 'TOTALP', 'BDEAD', 'TERMS', 'LOUTCOME', 'WEEKS', 'RACEMOM', 'RACEDAD', 'CIGNUM', 'DRINKNUM', 'ANEMIA', 'CARDIAC', 'ACLUNG', 'DIABETES', 'HERPES', 'HYDRAM', 'HEMOGLOB', 'HYPERCH', 'HYPERPR', 'ECLAMP', 'CERVIX', 'PINFANT', 'PRETERM', 'RENAL', 'RHSEN', 'UTERINE']\n",
      "qualitative : ['HISPMOM', 'HISPDAD']\n"
     ]
    }
   ],
   "source": [
    "#Identifying the categorical values from the training dataset\n",
    "quantitative = [f for f in Atrain1.columns if Atrain1.dtypes[f] != 'object']\n",
    "quantitative.remove('BWEIGHT')\n",
    "quantitative.remove('ID')\n",
    "qualitative = [f for f in Atrain1.columns if Atrain1.dtypes[f] == 'object']\n",
    "print('quantitative :',quantitative)\n",
    "print('qualitative :',qualitative)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 825,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0         M\n",
      "1         N\n",
      "2         N\n",
      "3         N\n",
      "4         N\n",
      "5         N\n",
      "6         M\n",
      "7         M\n",
      "8         N\n",
      "9         N\n",
      "10        N\n",
      "11        N\n",
      "12        N\n",
      "13        S\n",
      "14        N\n",
      "15        S\n",
      "16        N\n",
      "17        N\n",
      "18        N\n",
      "19        N\n",
      "20        M\n",
      "21        N\n",
      "22        N\n",
      "23        N\n",
      "24        N\n",
      "25        N\n",
      "26        N\n",
      "27        N\n",
      "28        N\n",
      "29        N\n",
      "         ..\n",
      "101370    N\n",
      "101371    N\n",
      "101372    N\n",
      "101373    N\n",
      "101374    M\n",
      "101375    M\n",
      "101376    N\n",
      "101377    N\n",
      "101378    M\n",
      "101379    P\n",
      "101380    N\n",
      "101381    N\n",
      "101382    N\n",
      "101383    N\n",
      "101384    N\n",
      "101385    N\n",
      "101386    N\n",
      "101387    N\n",
      "101388    P\n",
      "101389    N\n",
      "101390    N\n",
      "101391    S\n",
      "101392    S\n",
      "101393    N\n",
      "101394    M\n",
      "101395    M\n",
      "101396    N\n",
      "101397    N\n",
      "101398    N\n",
      "101399    N\n",
      "Name: HISPMOM, Length: 101400, dtype: category\n",
      "Categories (7, object): [C, M, N, O, P, S, U]\n",
      "0         M\n",
      "1         N\n",
      "2         N\n",
      "3         N\n",
      "4         N\n",
      "5         N\n",
      "6         M\n",
      "7         N\n",
      "8         N\n",
      "9         N\n",
      "10        N\n",
      "11        N\n",
      "12        N\n",
      "13        S\n",
      "14        N\n",
      "15        S\n",
      "16        N\n",
      "17        N\n",
      "18        N\n",
      "19        N\n",
      "20        M\n",
      "21        N\n",
      "22        N\n",
      "23        N\n",
      "24        N\n",
      "25        N\n",
      "26        N\n",
      "27        N\n",
      "28        N\n",
      "29        N\n",
      "         ..\n",
      "101370    N\n",
      "101371    N\n",
      "101372    N\n",
      "101373    N\n",
      "101374    N\n",
      "101375    M\n",
      "101376    N\n",
      "101377    N\n",
      "101378    M\n",
      "101379    N\n",
      "101380    N\n",
      "101381    N\n",
      "101382    N\n",
      "101383    N\n",
      "101384    N\n",
      "101385    N\n",
      "101386    N\n",
      "101387    N\n",
      "101388    M\n",
      "101389    N\n",
      "101390    N\n",
      "101391    P\n",
      "101392    N\n",
      "101393    N\n",
      "101394    M\n",
      "101395    M\n",
      "101396    N\n",
      "101397    N\n",
      "101398    N\n",
      "101399    N\n",
      "Name: HISPDAD, Length: 101400, dtype: category\n",
      "Categories (7, object): [C, M, N, O, P, S, U]\n"
     ]
    }
   ],
   "source": [
    "#Replace the missing qualitative variables\n",
    "for c in qualitative:\n",
    "    Atrain1[c] = Atrain1[c].astype('category')\n",
    "    print(Atrain1[c])\n",
    "    if Atrain1[c].isnull().any():\n",
    "        Atrain1[c] = Atrain1[c].cat.add_categories(['MISSING'])\n",
    "        Atrain1[c] = Atrain1[c].fillna('MISSING')\n",
    "        print(Atrain1[c])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 826,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['HISPMOM_E', 'HISPDAD_E']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "list"
      ]
     },
     "execution_count": 826,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Encode the qualitative variable as HISPMOM_E and HISPDAD_E\n",
    "def encode(frame, feature):\n",
    "    ordering = pd.DataFrame()\n",
    "    ordering['val'] = frame[feature].unique()\n",
    "    ordering.index = ordering.val\n",
    "    ordering['spmean'] = frame[[feature, 'BWEIGHT']].groupby(feature).mean()['BWEIGHT']\n",
    "    ordering = ordering.sort_values('spmean')\n",
    "    ordering['ordering'] = range(1, ordering.shape[0]+1)\n",
    "    ordering = ordering['ordering'].to_dict()\n",
    "    for cat, o in ordering.items():\n",
    "        frame.loc[frame[feature] == cat, feature+'_E'] = o\n",
    "qual_encoded = []\n",
    "for q in qualitative:  \n",
    "    encode(Atrain1, q)\n",
    "    qual_encoded.append(q+'_E')\n",
    "print(qual_encoded)\n",
    "type(qualitative)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 827,
   "metadata": {},
   "outputs": [],
   "source": [
    "Feature = quantitative + qual_encoded\n",
    "Upd_Dset = Atrain1[Feature]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 828,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x594 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Spearman's rank correlation\n",
    "def spearman(frame, Feature):\n",
    "    spr = pd.DataFrame()\n",
    "    spr['feature'] = Feature\n",
    "    spr['spearman'] = [frame[f].corr(frame['BWEIGHT'], 'spearman') for f in Feature]\n",
    "    spr = spr.sort_values('spearman')\n",
    "    plt.figure(figsize=(6, 0.25*len(Feature)))\n",
    "    sns.barplot(data=spr, y='feature', x='spearman', orient='h')\n",
    "    return \n",
    "Feature = quantitative \n",
    "spearman(Atrain1, Feature)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 829,
   "metadata": {},
   "outputs": [],
   "source": [
    "Correlated_Feature = [\"WEEKS\",\"GAINED\",\"RACEMOM\",\"RACEDAD\",\"VISITS\",\"MAGE\"]\n",
    "Upd_Dset = Upd_Dset[Correlated_Feature]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 3: Train and Test dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 830,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "       BWEIGHT\n",
      "1072    8.8125\n",
      "65946   6.0000\n",
      "70393   6.7500\n",
      "2599   10.4375\n",
      "67617   7.3750\n",
      "32787   6.0000\n",
      "10894   7.4375\n",
      "95293   7.1875\n",
      "23073   8.3125\n",
      "42819   7.6250\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(81120, 6)"
      ]
     },
     "execution_count": 830,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Split the dataset into 80% training and 20% testing\n",
    "from sklearn.model_selection import train_test_split\n",
    "ATrain, ATest = train_test_split(Upd_Dset, test_size=0.2, random_state=12345)\n",
    "BTrain, BTest = train_test_split(Atrain1[[\"BWEIGHT\"]], test_size=0.2, random_state=12345)\n",
    "print(BTest.head(10))\n",
    "ATrain.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 831,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 0.10582437 -0.6065972  -0.33815318 -0.35320679  3.09653639  1.89333778]\n",
      " [ 0.90245325 -0.38646079 -0.33815318 -0.35320679 -0.65010304  0.54925152]\n",
      " [ 0.10582437 -0.09294558 -0.33815318 -0.35320679 -0.11486883  0.04521917]\n",
      " ...\n",
      " [ 1.69908213 -0.01956678 -0.33815318 -0.35320679  2.02606798 -0.29080239]\n",
      " [ 0.10582437 -0.38646079 -0.33815318 -0.35320679 -0.11486883 -0.79483474]\n",
      " [-1.48743339  0.86097886 -0.33815318 -0.35320679 -0.11486883  1.725327  ]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(20280, 6)"
      ]
     },
     "execution_count": 831,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Applying a normalization technique to Training and Testing dataset\n",
    "stdscaler = StandardScaler()\n",
    "ATrain_X = stdscaler.fit_transform(ATrain)\n",
    "print(ATrain_X)\n",
    "\n",
    "\n",
    "stdscaler = StandardScaler()\n",
    "ATest_X = stdscaler.fit_transform(ATest)\n",
    "ATest_X.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 4: Model Creation using Linear Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 832,
   "metadata": {},
   "outputs": [],
   "source": [
    "ATrain_X = np.c_[np.ones(ATrain_X.shape[0]), ATrain_X]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 833,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(15) \n",
    "theta = np.random.rand(7, 1) \n",
    "m =  BTrain.size\n",
    "def linear_reg(x, y, theta):\n",
    "    prev_costs = []\n",
    "    prev_thetas = [theta]\n",
    "    for i in range(1000):\n",
    "        prediction = np.dot(x, theta)\n",
    "        error = prediction - y\n",
    "        cost = 1/(2*m) * np.dot(error.T, error)\n",
    "        prev_costs.append(cost)\n",
    "        theta = theta - (0.01 * (1/m) * np.dot(x.T, error))\n",
    "        prev_thetas.append(theta)\n",
    "    return prev_thetas, past_costs\n",
    "prev_thetas, prev_costs = linear_reg(ATrain_X, BTrain, theta)\n",
    "theta = prev_thetas[-1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 5: Birth Weight Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 834,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[7.27866234]\n",
      " [7.09491933]\n",
      " [7.39553163]\n",
      " ...\n",
      " [8.74971844]\n",
      " [7.14400035]\n",
      " [7.71908532]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 834,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ATest_X = np.c_[np.ones(ATest_X.shape[0]), ATest_X] \n",
    "Bpredict = np.dot(ATest_X, theta)\n",
    "print(Bpredict)\n",
    "BTest.ndim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 835,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BWEIGHT    1.070844\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "def error(actual, predicted):\n",
    "    return np.sqrt(np.sum(np.square(actual-predicted))/len(actual))\n",
    "print(error(BTest, Bpredict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 836,
   "metadata": {},
   "outputs": [],
   "source": [
    "Weight1 = error(BTest, Bpredict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 837,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean Absolute Error: 0.8380528960447409\n",
      "Mean Squared Error: 1.1467069837601163\n",
      "Root Mean Squared Error: 1.0708440520263052\n"
     ]
    }
   ],
   "source": [
    "print('Mean Absolute Error:', metrics.mean_absolute_error(BTest, Bpredict))  \n",
    "print('Mean Squared Error:', metrics.mean_squared_error(BTest, Bpredict))  \n",
    "print('Root Mean Squared Error:', np.sqrt(metrics.mean_squared_error(BTest, Bpredict)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 838,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BWEIGHT    1.070855\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "#Running this multiple times and calculate the Average RMSE \n",
    "np.random.seed(5)\n",
    "theta = np.random.rand(7, 1) \n",
    "m =  BTrain.size\n",
    "def linear_reg(x, y, theta):\n",
    "    prev_costs = []\n",
    "    prev_thetas = [theta]\n",
    "    for i in range(1000):\n",
    "        prediction = np.dot(x, theta)\n",
    "        error = prediction - y\n",
    "        cost = 1/(2*m) * np.dot(error.T, error)\n",
    "        prev_costs.append(cost)\n",
    "        theta = theta - (0.01 * (1/m) * np.dot(x.T, error))\n",
    "        prev_thetas.append(theta)\n",
    "    return prev_thetas, past_costs\n",
    "prev_thetas, prev_costs = linear_reg(ATrain_X, BTrain, theta)\n",
    "theta = prev_thetas[-1]\n",
    "Bpredict = np.dot(ATest_X, theta)\n",
    "Weight2 = error(BTest, Bpredict)\n",
    "print(Weight2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 839,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BWEIGHT    1.070838\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(8) \n",
    "theta = np.random.rand(7, 1) \n",
    "m =  BTrain.size\n",
    "def linear_reg(x, y, theta):\n",
    "    prev_costs = []\n",
    "    prev_thetas = [theta]\n",
    "    for i in range(1000):\n",
    "        prediction = np.dot(x, theta)\n",
    "        error = prediction - y\n",
    "        cost = 1/(2*m) * np.dot(error.T, error)\n",
    "        prev_costs.append(cost)\n",
    "        theta = theta - (0.01 * (1/m) * np.dot(x.T, error))\n",
    "        prev_thetas.append(theta)\n",
    "    return prev_thetas, past_costs\n",
    "prev_thetas, prev_costs = linear_reg(ATrain_X, BTrain, theta)\n",
    "theta = prev_thetas[-1]\n",
    "Bpredict = np.dot(ATest_X, theta)\n",
    "Weight3 = error(BTest, Bpredict)\n",
    "print(Weight3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 840,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BWEIGHT    1.071185\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(18) \n",
    "theta = np.random.rand(7, 1) \n",
    "m =  BTrain.size\n",
    "def linear_reg(x, y, theta):\n",
    "    prev_costs = []\n",
    "    prev_thetas = [theta]\n",
    "    for i in range(1000):\n",
    "        prediction = np.dot(x, theta)\n",
    "        error = prediction - y\n",
    "        cost = 1/(2*m) * np.dot(error.T, error)\n",
    "        prev_costs.append(cost)\n",
    "        theta = theta - (0.01 * (1/m) * np.dot(x.T, error))\n",
    "        prev_thetas.append(theta)\n",
    "    return prev_thetas, past_costs\n",
    "prev_thetas, prev_costs = linear_reg(ATrain_X, BTrain, theta)\n",
    "theta = prev_thetas[-1]\n",
    "Bpredict = np.dot(ATest_X, theta)\n",
    "Weight4 = error(BTest, Bpredict)\n",
    "print(Weight4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 841,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BWEIGHT    1.070965\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(16) \n",
    "theta = np.random.rand(7, 1) \n",
    "m =  BTrain.size\n",
    "def linear_reg(x, y, theta):\n",
    "    prev_costs = []\n",
    "    prev_thetas = [theta]\n",
    "    for i in range(1000):\n",
    "        prediction = np.dot(x, theta)\n",
    "        error = prediction - y\n",
    "        cost = 1/(2*m) * np.dot(error.T, error)\n",
    "        prev_costs.append(cost)\n",
    "        theta = theta - (0.01 * (1/m) * np.dot(x.T, error))\n",
    "        prev_thetas.append(theta)\n",
    "    return prev_thetas, past_costs\n",
    "prev_thetas, prev_costs = linear_reg(ATrain_X, BTrain, theta)\n",
    "theta = prev_thetas[-1]\n",
    "Bpredict = np.dot(ATest_X, theta)\n",
    "Weight5 = error(BTest, Bpredict)\n",
    "print(Weight5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 842,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AverageRMSE : BWEIGHT    1.070937\n",
      "dtype: float64\n",
      "StdDev : 2.645803115690209e-05\n",
      "RMSE : BWEIGHT    1.070911\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "AverageRMSE = (Weight1 + Weight2  + Weight3 + Weight4 + Weight5 )/ 5\n",
    "StdDev = np.std([Weight1, Weight2 , Weight3 , Weight4 , Weight5])/5\n",
    "RMSE = AverageRMSE +- StdDev\n",
    "print('AverageRMSE :',AverageRMSE)\n",
    "print('StdDev :',StdDev)\n",
    "print('RMSE :',RMSE)"
   ]
  }
 ],
 "metadata": {
  "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}