ex61.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import statsmodels.api as sm\n",
    "import statsmodels.formula.api as smf\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy import stats\n",
    "from IPython.display import Markdown as md\n",
    "from IPython.core.display import display, HTML\n",
    "np.random.seed(42)\n",
    "sns.set()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>sex</th>\n",
       "      <th>couples</th>\n",
       "      <th>women_alone</th>\n",
       "      <th>bs_hiv</th>\n",
       "      <th>bupacts</th>\n",
       "      <th>fupacts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>434</td>\n",
       "      <td>434.000000</td>\n",
       "      <td>434.000000</td>\n",
       "      <td>434</td>\n",
       "      <td>434.000000</td>\n",
       "      <td>434.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>man</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>negative</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>217</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>337</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.373272</td>\n",
       "      <td>0.336406</td>\n",
       "      <td>NaN</td>\n",
       "      <td>25.910139</td>\n",
       "      <td>16.489580</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.484232</td>\n",
       "      <td>0.473025</td>\n",
       "      <td>NaN</td>\n",
       "      <td>31.917963</td>\n",
       "      <td>26.825769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>5.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>36.000000</td>\n",
       "      <td>20.925600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>300.000000</td>\n",
       "      <td>200.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        sex     couples  women_alone    bs_hiv     bupacts     fupacts\n",
       "count   434  434.000000   434.000000       434  434.000000  434.000000\n",
       "unique    2         NaN          NaN         2         NaN         NaN\n",
       "top     man         NaN          NaN  negative         NaN         NaN\n",
       "freq    217         NaN          NaN       337         NaN         NaN\n",
       "mean    NaN    0.373272     0.336406       NaN   25.910139   16.489580\n",
       "std     NaN    0.484232     0.473025       NaN   31.917963   26.825769\n",
       "min     NaN    0.000000     0.000000       NaN    0.000000    0.000000\n",
       "25%     NaN    0.000000     0.000000       NaN    5.000000    0.000000\n",
       "50%     NaN    0.000000     0.000000       NaN   15.000000    5.000000\n",
       "75%     NaN    1.000000     1.000000       NaN   36.000000   20.925600\n",
       "max     NaN    1.000000     1.000000       NaN  300.000000  200.000000"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hiv_df = pd.read_stata(\"./ARM_Data_extra/examples/risky.behavior/risky_behaviors.dta\")\n",
    "hiv_df.describe(include='all')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 16.423400\n",
      "         Iterations 5\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>Poisson Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>        <td>fupacts</td>     <th>  No. Observations:  </th>  <td>   434</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                <td>Poisson</td>     <th>  Df Residuals:      </th>  <td>   431</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>                 <td>MLE</td>       <th>  Df Model:          </th>  <td>     2</td>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>            <td>Thu, 26 Mar 2020</td> <th>  Pseudo R-squ.:     </th>  <td>0.02566</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                <td>16:12:51</td>     <th>  Log-Likelihood:    </th> <td> -7127.8</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>converged:</th>             <td>True</td>       <th>  LL-Null:           </th> <td> -7315.5</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>     <td>nonrobust</td>    <th>  LLR p-value:       </th> <td>2.986e-82</td>\n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "       <td></td>          <th>coef</th>     <th>std err</th>      <th>z</th>      <th>P>|z|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Intercept</th>   <td>    3.0902</td> <td>    0.019</td> <td>  162.635</td> <td> 0.000</td> <td>    3.053</td> <td>    3.127</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>couples</th>     <td>   -0.3226</td> <td>    0.027</td> <td>  -11.791</td> <td> 0.000</td> <td>   -0.376</td> <td>   -0.269</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>women_alone</th> <td>   -0.5741</td> <td>    0.030</td> <td>  -18.987</td> <td> 0.000</td> <td>   -0.633</td> <td>   -0.515</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                          Poisson Regression Results                          \n",
       "==============================================================================\n",
       "Dep. Variable:                fupacts   No. Observations:                  434\n",
       "Model:                        Poisson   Df Residuals:                      431\n",
       "Method:                           MLE   Df Model:                            2\n",
       "Date:                Thu, 26 Mar 2020   Pseudo R-squ.:                 0.02566\n",
       "Time:                        16:12:51   Log-Likelihood:                -7127.8\n",
       "converged:                       True   LL-Null:                       -7315.5\n",
       "Covariance Type:            nonrobust   LLR p-value:                 2.986e-82\n",
       "===============================================================================\n",
       "                  coef    std err          z      P>|z|      [0.025      0.975]\n",
       "-------------------------------------------------------------------------------\n",
       "Intercept       3.0902      0.019    162.635      0.000       3.053       3.127\n",
       "couples        -0.3226      0.027    -11.791      0.000      -0.376      -0.269\n",
       "women_alone    -0.5741      0.030    -18.987      0.000      -0.633      -0.515\n",
       "===============================================================================\n",
       "\"\"\""
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mdl_a = smf.poisson(data=hiv_df, formula=\"fupacts ~ couples + women_alone\").fit()\n",
    "mdl_a.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Overdispersion ratio is 44.04779229512923, p-value of it is 1.0'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "yhat = mdl_a.predict(hiv_df)\n",
    "resid = hiv_df.fupacts - yhat # mdl_a.resid could be used here as well, going more straightforward for transparency\n",
    "std_residuals = resid/np.sqrt(yhat)\n",
    "sr_sq_sum = np.sum(std_residuals*std_residuals)\n",
    "mdl_a_overdispersion_ratio = sr_sq_sum/(len(hiv_df) - 2) # 2 for the number of predictors\n",
    "mdl_a_overdispersion_p_value = stats.chi2.cdf(sr_sq_sum, df=mdl_a.df_resid)\n",
    "f\"Overdispersion ratio is {mdl_a_overdispersion_ratio}, p-value of it is {mdl_a_overdispersion_p_value}\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x864 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "gs = fig.add_gridspec(2,2)\n",
    "ax1 = fig.add_subplot(gs[0, 0])\n",
    "ax2 = fig.add_subplot(gs[1, 0])\n",
    "ax3 = fig.add_subplot(gs[:, 1])\n",
    "sns.scatterplot(x=mdl_a.predict(hiv_df), y=resid, ax=ax1)\n",
    "sns.scatterplot(x=mdl_a.predict(hiv_df), y=std_residuals, ax=ax2)\n",
    "sns.distplot(std_residuals, kde=True, axlabel=\"Standardized residuals\")\n",
    "ax1.set_xlabel(\"Prediction\")\n",
    "ax2.set_xlabel(\"Prediction\")\n",
    "ax1.set_ylabel(\"Residuals\")\n",
    "ax2.set_ylabel(\"Standardized resudials\")\n",
    "fig.set_size_inches(16, 12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Bad fit, the overdispersion factor is very high."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 13.282019\n",
      "         Iterations 9\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>Poisson Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>        <td>fupacts</td>     <th>  No. Observations:  </th>  <td>   434</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                <td>Poisson</td>     <th>  Df Residuals:      </th>  <td>   428</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>                 <td>MLE</td>       <th>  Df Model:          </th>  <td>     5</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>            <td>Thu, 26 Mar 2020</td> <th>  Pseudo R-squ.:     </th>  <td>0.2120</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                <td>16:12:52</td>     <th>  Log-Likelihood:    </th> <td> -5764.4</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>converged:</th>             <td>True</td>       <th>  LL-Null:           </th> <td> -7315.5</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>     <td>nonrobust</td>    <th>  LLR p-value:       </th>  <td> 0.000</td> \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "           <td></td>             <th>coef</th>     <th>std err</th>      <th>z</th>      <th>P>|z|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Intercept</th>          <td>    2.8961</td> <td>    0.023</td> <td>  124.821</td> <td> 0.000</td> <td>    2.851</td> <td>    2.942</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sex[T.man]</th>         <td>   -0.1085</td> <td>    0.024</td> <td>   -4.572</td> <td> 0.000</td> <td>   -0.155</td> <td>   -0.062</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>bs_hiv[T.positive]</th> <td>   -0.4376</td> <td>    0.035</td> <td>  -12.371</td> <td> 0.000</td> <td>   -0.507</td> <td>   -0.368</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>couples</th>            <td>   -0.4104</td> <td>    0.028</td> <td>  -14.541</td> <td> 0.000</td> <td>   -0.466</td> <td>   -0.355</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>women_alone</th>        <td>   -0.6643</td> <td>    0.031</td> <td>  -21.496</td> <td> 0.000</td> <td>   -0.725</td> <td>   -0.604</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>bupacts</th>            <td>    0.0108</td> <td>    0.000</td> <td>   62.043</td> <td> 0.000</td> <td>    0.010</td> <td>    0.011</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                          Poisson Regression Results                          \n",
       "==============================================================================\n",
       "Dep. Variable:                fupacts   No. Observations:                  434\n",
       "Model:                        Poisson   Df Residuals:                      428\n",
       "Method:                           MLE   Df Model:                            5\n",
       "Date:                Thu, 26 Mar 2020   Pseudo R-squ.:                  0.2120\n",
       "Time:                        16:12:52   Log-Likelihood:                -5764.4\n",
       "converged:                       True   LL-Null:                       -7315.5\n",
       "Covariance Type:            nonrobust   LLR p-value:                     0.000\n",
       "======================================================================================\n",
       "                         coef    std err          z      P>|z|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------------\n",
       "Intercept              2.8961      0.023    124.821      0.000       2.851       2.942\n",
       "sex[T.man]            -0.1085      0.024     -4.572      0.000      -0.155      -0.062\n",
       "bs_hiv[T.positive]    -0.4376      0.035    -12.371      0.000      -0.507      -0.368\n",
       "couples               -0.4104      0.028    -14.541      0.000      -0.466      -0.355\n",
       "women_alone           -0.6643      0.031    -21.496      0.000      -0.725      -0.604\n",
       "bupacts                0.0108      0.000     62.043      0.000       0.010       0.011\n",
       "======================================================================================\n",
       "\"\"\""
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mdl_b = smf.poisson(data=hiv_df, formula=\"fupacts ~ couples + women_alone + sex + bs_hiv + bupacts\").fit()\n",
    "mdl_b.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Overdispersion ratio is 29.942248987890537, p-value of it is 1.0'"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "yhat = mdl_b.predict(hiv_df)\n",
    "resid = hiv_df.fupacts - yhat # mdl_a.resid could be used here as well, going more straightforward for transparency\n",
    "std_residuals = resid/np.sqrt(yhat)\n",
    "sr_sq_sum = np.sum(std_residuals*std_residuals)\n",
    "mdl_b_overdispersion_ratio = sr_sq_sum/(len(hiv_df) - 5) # 5 for the number of predictors\n",
    "mdl_b_overdispersion_p_value = stats.chi2.cdf(sr_sq_sum, df=mdl_b.df_resid)\n",
    "f\"Overdispersion ratio is {mdl_b_overdispersion_ratio}, p-value of it is {mdl_b_overdispersion_p_value}\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Better fit, the overdispersion went down a bit, yet is still too high."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we have to dive into depths of github as Quasi Poisson model is currently not implemented in statsmodel. Yet there's a [py-glm](https://github.com/madrury/py-glm) package which should work"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from glm.glm import GLM\n",
    "from glm.families import QuasiPoisson, Poisson"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "QuasiPoisson GLM Model Summary.\n",
      "===============================================\n",
      "Name         Parameter Estimate  Standard Error\n",
      "-----------------------------------------------\n",
      "Intercept                  2.90            0.11\n",
      "sex[T.man]                -0.11            0.12\n",
      "bs_hiv[T.positive]        -0.44            0.17\n",
      "couples                   -0.41            0.14\n",
      "women_alone               -0.66            0.15\n",
      "bupacts                    0.01            0.00\n"
     ]
    }
   ],
   "source": [
    "mdl_c = GLM(family=QuasiPoisson())\n",
    "mdl_c.fit(X=hiv_df, formula=\"fupacts ~ couples + women_alone + sex + bs_hiv + bupacts\")\n",
    "mdl_c.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(10204.40440327)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mdl_c.deviance_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(23.84206636)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mdl_c.dispersion_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As suggested by these [two](https://github.com/statsmodels/statsmodels/issues/2804) [links](https://github.com/statsmodels/statsmodels/issues/4942) we try an hc0 covariance type fit. The covariance types description is present [here](https://cran.r-project.org/web/packages/sandwich/vignettes/sandwich.pdf)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 13.282019\n",
      "         Iterations 9\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<caption>Poisson Regression Results</caption>\n",
       "<tr>\n",
       "  <th>Dep. Variable:</th>        <td>fupacts</td>     <th>  No. Observations:  </th>  <td>   434</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Model:</th>                <td>Poisson</td>     <th>  Df Residuals:      </th>  <td>   428</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Method:</th>                 <td>MLE</td>       <th>  Df Model:          </th>  <td>     5</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Date:</th>            <td>Thu, 26 Mar 2020</td> <th>  Pseudo R-squ.:     </th>  <td>0.2120</td> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Time:</th>                <td>16:12:52</td>     <th>  Log-Likelihood:    </th> <td> -5764.4</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>converged:</th>             <td>True</td>       <th>  LL-Null:           </th> <td> -7315.5</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Covariance Type:</th>        <td>HC0</td>       <th>  LLR p-value:       </th>  <td> 0.000</td> \n",
       "</tr>\n",
       "</table>\n",
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "           <td></td>             <th>coef</th>     <th>std err</th>      <th>z</th>      <th>P>|z|</th>  <th>[0.025</th>    <th>0.975]</th>  \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Intercept</th>          <td>    2.8961</td> <td>    0.129</td> <td>   22.479</td> <td> 0.000</td> <td>    2.644</td> <td>    3.149</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>sex[T.man]</th>         <td>   -0.1085</td> <td>    0.143</td> <td>   -0.758</td> <td> 0.448</td> <td>   -0.389</td> <td>    0.172</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>bs_hiv[T.positive]</th> <td>   -0.4376</td> <td>    0.200</td> <td>   -2.186</td> <td> 0.029</td> <td>   -0.830</td> <td>   -0.045</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>couples</th>            <td>   -0.4104</td> <td>    0.176</td> <td>   -2.329</td> <td> 0.020</td> <td>   -0.756</td> <td>   -0.065</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>women_alone</th>        <td>   -0.6643</td> <td>    0.151</td> <td>   -4.395</td> <td> 0.000</td> <td>   -0.961</td> <td>   -0.368</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>bupacts</th>            <td>    0.0108</td> <td>    0.001</td> <td>    8.376</td> <td> 0.000</td> <td>    0.008</td> <td>    0.013</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.summary.Summary'>\n",
       "\"\"\"\n",
       "                          Poisson Regression Results                          \n",
       "==============================================================================\n",
       "Dep. Variable:                fupacts   No. Observations:                  434\n",
       "Model:                        Poisson   Df Residuals:                      428\n",
       "Method:                           MLE   Df Model:                            5\n",
       "Date:                Thu, 26 Mar 2020   Pseudo R-squ.:                  0.2120\n",
       "Time:                        16:12:52   Log-Likelihood:                -5764.4\n",
       "converged:                       True   LL-Null:                       -7315.5\n",
       "Covariance Type:                  HC0   LLR p-value:                     0.000\n",
       "======================================================================================\n",
       "                         coef    std err          z      P>|z|      [0.025      0.975]\n",
       "--------------------------------------------------------------------------------------\n",
       "Intercept              2.8961      0.129     22.479      0.000       2.644       3.149\n",
       "sex[T.man]            -0.1085      0.143     -0.758      0.448      -0.389       0.172\n",
       "bs_hiv[T.positive]    -0.4376      0.200     -2.186      0.029      -0.830      -0.045\n",
       "couples               -0.4104      0.176     -2.329      0.020      -0.756      -0.065\n",
       "women_alone           -0.6643      0.151     -4.395      0.000      -0.961      -0.368\n",
       "bupacts                0.0108      0.001      8.376      0.000       0.008       0.013\n",
       "======================================================================================\n",
       "\"\"\""
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mdl_c1 = smf.poisson(data=hiv_df, formula=\"fupacts ~ couples + women_alone + sex + bs_hiv + bupacts\").fit(cov_type='HC0')\n",
    "mdl_c1.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Overdispersion ratio is 29.942248987890537, p-value of it is 1.0'"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "yhat = mdl_c1.predict(hiv_df)\n",
    "resid = hiv_df.fupacts - yhat # mdl_a.resid could be used here as well, going more straightforward for transparency\n",
    "std_residuals = resid/np.sqrt(yhat)\n",
    "sr_sq_sum = np.sum(std_residuals*std_residuals)\n",
    "mdl_c1_overdispersion_ratio = sr_sq_sum/(len(hiv_df) - 5) # 5 for the number of predictors\n",
    "mdl_c1_overdispersion_p_value = stats.chi2.cdf(sr_sq_sum, df=mdl_c1.df_resid)\n",
    "f\"Overdispersion ratio is {mdl_c1_overdispersion_ratio}, p-value of it is {mdl_c1_overdispersion_p_value}\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x864 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "gs = fig.add_gridspec(2,2)\n",
    "ax1 = fig.add_subplot(gs[0, 0])\n",
    "ax2 = fig.add_subplot(gs[1, 0])\n",
    "ax3 = fig.add_subplot(gs[:, 1])\n",
    "sns.scatterplot(x=mdl_c1.predict(hiv_df), y=resid, ax=ax1)\n",
    "sns.scatterplot(x=mdl_c1.predict(hiv_df), y=std_residuals, ax=ax2)\n",
    "sns.distplot(std_residuals, kde=True, axlabel=\"Standardized residuals\")\n",
    "ax1.set_xlabel(\"Prediction\")\n",
    "ax2.set_xlabel(\"Prediction\")\n",
    "ax1.set_ylabel(\"Residuals\")\n",
    "ax2.set_ylabel(\"Standardized resudials\")\n",
    "fig.set_size_inches(16, 12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The fit got better, from the standard error values it is clear that the overdispersion is considered in the fit process.\n",
    "\n",
    "Coefficient interpretation-wise it is clear that the councelling provides a positive effect, yet the effect is not that strong when the treatment is done for a couple rather than a women."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": 4
}