{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cvxpy as cvx\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "from itertools import product\n",
    "from sklearn import model_selection as ms\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data pre-processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read in the *Child Health and Development Studies* data from the R package [`mosaicData`](https://cran.r-project.org/package=mosaicData)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "gestation = pd.read_csv('https://vincentarelbundock.github.io/Rdatasets/csv/mosaicData/Gestation.csv', index_col=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop some columns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "gestation.drop(columns=['id', 'pluralty', 'outcome', 'date'], inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check for missing values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dwt          0.403722\n",
       "dht          0.398058\n",
       "inc          0.100324\n",
       "wt.1         0.029126\n",
       "drace        0.025081\n",
       "ht           0.017799\n",
       "number       0.016990\n",
       "ded          0.010518\n",
       "gestation    0.010518\n",
       "time         0.008091\n",
       "smoke        0.008091\n",
       "dage         0.005663\n",
       "age          0.001618\n",
       "ed           0.000809\n",
       "race         0.000809\n",
       "marital      0.000000\n",
       "parity       0.000000\n",
       "wt           0.000000\n",
       "sex          0.000000\n",
       "dtype: float64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gestation.isnull().mean().sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop variables with many missing values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "gestation.drop(columns=['dht', 'dwt', 'inc'], inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Drop observations with missing values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "gestation.dropna(inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create dummies for categorical variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "gestation = pd.get_dummies(gestation, columns=['race', 'ed', 'drace', 'ded', 'marital', 'smoke', 'time', 'number'], drop_first=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>gestation</th>\n",
       "      <th>sex</th>\n",
       "      <th>wt</th>\n",
       "      <th>parity</th>\n",
       "      <th>age</th>\n",
       "      <th>ht</th>\n",
       "      <th>wt.1</th>\n",
       "      <th>dage</th>\n",
       "      <th>race_1.0</th>\n",
       "      <th>race_2.0</th>\n",
       "      <th>...</th>\n",
       "      <th>time_8.0</th>\n",
       "      <th>time_9.0</th>\n",
       "      <th>number_1.0</th>\n",
       "      <th>number_2.0</th>\n",
       "      <th>number_3.0</th>\n",
       "      <th>number_4.0</th>\n",
       "      <th>number_5.0</th>\n",
       "      <th>number_6.0</th>\n",
       "      <th>number_7.0</th>\n",
       "      <th>number_8.0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>284.0</td>\n",
       "      <td>1</td>\n",
       "      <td>120</td>\n",
       "      <td>1</td>\n",
       "      <td>27.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>282.0</td>\n",
       "      <td>1</td>\n",
       "      <td>113</td>\n",
       "      <td>2</td>\n",
       "      <td>33.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>135.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>279.0</td>\n",
       "      <td>1</td>\n",
       "      <td>128</td>\n",
       "      <td>1</td>\n",
       "      <td>28.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>115.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>282.0</td>\n",
       "      <td>1</td>\n",
       "      <td>108</td>\n",
       "      <td>1</td>\n",
       "      <td>23.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>125.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>286.0</td>\n",
       "      <td>1</td>\n",
       "      <td>136</td>\n",
       "      <td>4</td>\n",
       "      <td>25.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>93.0</td>\n",
       "      <td>28.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   gestation  sex   wt  parity   age    ht   wt.1  dage  race_1.0  race_2.0  \\\n",
       "1      284.0    1  120       1  27.0  62.0  100.0  31.0         0         0   \n",
       "2      282.0    1  113       2  33.0  64.0  135.0  38.0         0         0   \n",
       "3      279.0    1  128       1  28.0  64.0  115.0  32.0         0         0   \n",
       "5      282.0    1  108       1  23.0  67.0  125.0  24.0         0         0   \n",
       "6      286.0    1  136       4  25.0  62.0   93.0  28.0         0         0   \n",
       "\n",
       "      ...      time_8.0  time_9.0  number_1.0  number_2.0  number_3.0  \\\n",
       "1     ...             0         0           0           0           0   \n",
       "2     ...             0         0           0           0           0   \n",
       "3     ...             0         0           1           0           0   \n",
       "5     ...             0         0           0           0           0   \n",
       "6     ...             0         0           0           1           0   \n",
       "\n",
       "   number_4.0  number_5.0  number_6.0  number_7.0  number_8.0  \n",
       "1           0           0           0           0           0  \n",
       "2           0           0           0           0           0  \n",
       "3           0           0           0           0           0  \n",
       "5           0           1           0           0           0  \n",
       "6           0           0           0           0           0  \n",
       "\n",
       "[5 rows x 65 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gestation.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Modelling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Split data into training and test sets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "gestation_train, gestation_test = ms.train_test_split(gestation, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Separate outcome from predictors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = gestation_train.drop(columns=['wt'])\n",
    "y_train = gestation_train['wt']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = gestation_test.drop(columns=['wt'])\n",
    "y_test = gestation_test['wt']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define the mean squared error as a function of the regression coefficients `betas` given `X` and `y`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def mse(betas, X, y):\n",
    "    return cvx.pnorm(cvx.matmul(X, betas) - y, p=2)**2 / X.shape[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define the elastic net penalty as a function of the regression coefficients `betas` given the strengths `lambda_l1` (for the lasso penalty) and `lambda_l2` (for the ridge penalty)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def elastic_net(betas, lambda_l1, lambda_l2):\n",
    "    return lambda_l1 * cvx.pnorm(betas, p=1) + \\\n",
    "           lambda_l2 * cvx.pnorm(betas, p=2)**2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define the regularised loss that will be minimised as the sum of `mse` and `elastic_net`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def regularized_loss(betas, lambda_l1, lambda_l2, X, y):\n",
    "    return mse(betas, X, y) + \\\n",
    "           elastic_net(betas, lambda_l1, lambda_l2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define a function to fit models given sequences of `lambdas` and `l1_ratios`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def fit(X_train, y_train, X_test, y_test, lambdas, l1_ratios):\n",
    "    betas = cvx.Variable(shape=X_train.shape[1])\n",
    "    lambda_l1 = cvx.Parameter(nonneg=True)\n",
    "    lambda_l2 = cvx.Parameter(nonneg=True)\n",
    "    problem = cvx.Problem(cvx.Minimize(regularized_loss(betas, lambda_l1, lambda_l2, X_train, y_train)))\n",
    "    results = []\n",
    "    for lambda_, l1_ratio in product(lambdas, l1_ratios):\n",
    "        lambda_l1.value = lambda_ * l1_ratio\n",
    "        lambda_l2.value = lambda_ * (1 - l1_ratio)\n",
    "        problem.solve()\n",
    "        results.append({\n",
    "            'lambda': lambda_,\n",
    "            'l1_ratio': l1_ratio,\n",
    "            'coefs': betas.value,\n",
    "            'mse_train': mse(betas, X_train, y_train).value,\n",
    "            'mse_test': mse(betas, X_test, y_test).value,\n",
    "        })\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_results = fit(X_train, y_train, X_test, y_test, np.logspace(-2, 2, 10), [0, 0.25, 0.5, 0.75, 1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Extract and plot the mean squared error in the test set as a function of `lambda` and `l1_ratio`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "cv_metrics = pd.DataFrame(cv_results).drop(columns=['coefs'])"
   ]
  },
  {
   "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>l1_ratio</th>\n",
       "      <th>lambda</th>\n",
       "      <th>mse_test</th>\n",
       "      <th>mse_train</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.01</td>\n",
       "      <td>262.517981</td>\n",
       "      <td>230.772366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.25</td>\n",
       "      <td>0.01</td>\n",
       "      <td>263.030008</td>\n",
       "      <td>230.047452</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.50</td>\n",
       "      <td>0.01</td>\n",
       "      <td>263.539374</td>\n",
       "      <td>229.085362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.75</td>\n",
       "      <td>0.01</td>\n",
       "      <td>263.961701</td>\n",
       "      <td>227.622517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.00</td>\n",
       "      <td>0.01</td>\n",
       "      <td>266.286687</td>\n",
       "      <td>224.393952</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   l1_ratio  lambda    mse_test   mse_train\n",
       "0      0.00    0.01  262.517981  230.772366\n",
       "1      0.25    0.01  263.030008  230.047452\n",
       "2      0.50    0.01  263.539374  229.085362\n",
       "3      0.75    0.01  263.961701  227.622517\n",
       "4      1.00    0.01  266.286687  224.393952"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_metrics.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fafbbae67b8>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fafbbae6c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8, 6))\n",
    "for label, df in cv_metrics.groupby('l1_ratio'):\n",
    "    df.plot(kind='line', x='lambda', y='mse_test', ax=ax, label=label, logx=True)\n",
    "plt.legend()"
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}