{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Introduction to Bayesian Regression Analysis" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We will review Frequentist regression first and then learn the Bayesian methods. This will help us understand the basic frameworks in both approaches. " ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "%matplotlib inline\n", "import seaborn as sns\n", "\n", "# packages for frequentist regression\n", "import statsmodels.api as sm\n", "# package for bayesian regression\n", "import pystan" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
crimzninduschasnoxrmagedisradtaxptratioblacklstatmedv
00.0063218.02.3100.5386.57565.24.0900129615.3396.904.9824.0
10.027310.07.0700.4696.42178.94.9671224217.8396.909.1421.6
20.027290.07.0700.4697.18561.14.9671224217.8392.834.0334.7
30.032370.02.1800.4586.99845.86.0622322218.7394.632.9433.4
40.069050.02.1800.4587.14754.26.0622322218.7396.905.3336.2
\n", "
" ], "text/plain": [ " crim zn indus chas nox rm age dis rad tax ptratio \\\n", "0 0.00632 18.0 2.31 0 0.538 6.575 65.2 4.0900 1 296 15.3 \n", "1 0.02731 0.0 7.07 0 0.469 6.421 78.9 4.9671 2 242 17.8 \n", "2 0.02729 0.0 7.07 0 0.469 7.185 61.1 4.9671 2 242 17.8 \n", "3 0.03237 0.0 2.18 0 0.458 6.998 45.8 6.0622 3 222 18.7 \n", "4 0.06905 0.0 2.18 0 0.458 7.147 54.2 6.0622 3 222 18.7 \n", "\n", " black lstat medv \n", "0 396.90 4.98 24.0 \n", "1 396.90 9.14 21.6 \n", "2 392.83 4.03 34.7 \n", "3 394.63 2.94 33.4 \n", "4 396.90 5.33 36.2 " ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# read .csv to Python dataframe\n", "# also can be loaded from sklearn.datasets\n", "boston_df = pd.read_csv(\"Boston.csv\") \n", "boston_df.head()" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Index(['crim', 'zn', 'indus', 'chas', 'nox', 'rm', 'age', 'dis', 'rad', 'tax',\n", " 'ptratio', 'black', 'lstat', 'medv'],\n", " dtype='object')" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "boston_df.columns" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Frequentist Linear Regression" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "It is always encouraged (actually, necessary) to conduct explatory data analysis before choosing a model. Here we will try a few scatter plots to determine if this data would fit for linear regression model. " ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'Median of owner-occupied homes ($1000s)')" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.scatter(x=boston_df['lstat'], y=boston_df['medv'])\n", "plt.xlabel('% of lower income population')\n", "plt.ylabel('Median of owner-occupied homes ($1000s)')" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'Median of owner-occupied homes ($1000s)')" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.scatter(x=boston_df['age'], y=boston_df['medv'])\n", "plt.xlabel('Age')\n", "plt.ylabel('Median of owner-occupied homes ($1000s)')" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "1.73\n", "37.97\n" ] } ], "source": [ "print(np.min(boston_df['lstat']))\n", "print(np.max(boston_df['lstat']))" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "
OLS Regression Results
Dep. Variable: medv R-squared: 0.551
Model: OLS Adj. R-squared: 0.549
Method: Least Squares F-statistic: 309.0
Date: Mon, 06 May 2019 Prob (F-statistic): 2.98e-88
Time: 14:37:09 Log-Likelihood: -1637.5
No. Observations: 506 AIC: 3281.
Df Residuals: 503 BIC: 3294.
Df Model: 2
Covariance Type: nonrobust
\n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "
coef std err t P>|t| [0.025 0.975]
age 0.0345 0.012 2.826 0.005 0.011 0.059
lstat -1.0321 0.048 -21.416 0.000 -1.127 -0.937
intercept 33.2228 0.731 45.458 0.000 31.787 34.659
\n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "\n", " \n", "\n", "
Omnibus: 124.288 Durbin-Watson: 0.945
Prob(Omnibus): 0.000 Jarque-Bera (JB): 244.026
Skew: 1.362 Prob(JB): 1.02e-53
Kurtosis: 5.038 Cond. No. 201.


Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified." ], "text/plain": [ "\n", "\"\"\"\n", " OLS Regression Results \n", "==============================================================================\n", "Dep. Variable: medv R-squared: 0.551\n", "Model: OLS Adj. R-squared: 0.549\n", "Method: Least Squares F-statistic: 309.0\n", "Date: Mon, 06 May 2019 Prob (F-statistic): 2.98e-88\n", "Time: 14:37:09 Log-Likelihood: -1637.5\n", "No. Observations: 506 AIC: 3281.\n", "Df Residuals: 503 BIC: 3294.\n", "Df Model: 2 \n", "Covariance Type: nonrobust \n", "==============================================================================\n", " coef std err t P>|t| [0.025 0.975]\n", "------------------------------------------------------------------------------\n", "age 0.0345 0.012 2.826 0.005 0.011 0.059\n", "lstat -1.0321 0.048 -21.416 0.000 -1.127 -0.937\n", "intercept 33.2228 0.731 45.458 0.000 31.787 34.659\n", "==============================================================================\n", "Omnibus: 124.288 Durbin-Watson: 0.945\n", "Prob(Omnibus): 0.000 Jarque-Bera (JB): 244.026\n", "Skew: 1.362 Prob(JB): 1.02e-53\n", "Kurtosis: 5.038 Cond. No. 201.\n", "==============================================================================\n", "\n", "Warnings:\n", "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n", "\"\"\"" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Running linear regression using statsmodels\n", "y = 'medv'\n", "x = ['age', 'lstat']\n", "# Create a design matrix (add one column with 1.0's)\n", "ones = np.ones(boston_df.shape[0])\n", "design_mat = boston_df[x].assign(intercept=ones)\n", "# Add a constant term like so:\n", "boston_fit = sm.OLS(boston_df[y], design_mat).fit()\n", "boston_fit.summary()" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'residuals')" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Calculate resiudals: fitted - observed \n", "residuals = boston_fit.fittedvalues - boston_df['medv']\n", "\n", "# Diagnostic plot: predictor vs. residuals\n", "fig = plt.figure(figsize=(12,5))\n", "ax1 = plt.subplot(1, 2, 1)\n", "ax2 = plt.subplot(1, 2, 2)\n", "ax1.scatter(boston_df['age'], residuals)\n", "ax1.set_title('age vs. residuals', fontsize=18)\n", "ax2.scatter(boston_df['lstat'], residuals)\n", "ax2.set_title('lstat vs. residuals', fontsize=18)\n", "ax1.set_xlabel('age', fontsize=18)\n", "ax1.set_ylabel('residuals', fontsize=18)\n", "ax2.set_xlabel('lstat', fontsize=18)\n", "ax2.set_ylabel('residuals', fontsize=18)" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'residual')" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# We can add a squared lstat\n", "lstat2 = np.square(design_mat['lstat'])\n", "design_mat2 = design_mat.assign(lstat2=lstat2)\n", "design_mat2\n", "boston_fit2 = sm.OLS(boston_df[y], design_mat2).fit()\n", "residuals2 = boston_fit2.fittedvalues - boston_df['medv']\n", "plt.scatter(boston_df['lstat'], residuals2)\n", "plt.title('lstat vs. residuals after adding lstat^2', fontsize=18)\n", "plt.xlabel('lstat', fontsize=18)\n", "plt.ylabel('residual', fontsize=18)" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'fitted medv')" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Diagnostics plot: observed vs. fitted \n", "fig = plt.figure(figsize=(12,5))\n", "ax1 = plt.subplot(1, 2, 1)\n", "ax2 = plt.subplot(1, 2, 2)\n", "ax1.scatter(boston_df['medv'], boston_fit.fittedvalues)\n", "ax1.set_title('medv ~ age + lstat', fontsize=18)\n", "ax2.scatter(boston_df['medv'], boston_fit2.fittedvalues)\n", "ax2.set_title('medv ~ age + lstat + lstat^2', fontsize=18)\n", "ax1.set_xlabel('medv', fontsize=18)\n", "ax1.set_ylabel('fitted medv', fontsize=18)\n", "ax2.set_xlabel('medv', fontsize=18)\n", "ax2.set_ylabel('fitted medv', fontsize=18)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Bayesian Linear Regression with PyStan" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Seminar slides on Bayesian Regression include the following example." ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "INFO:pystan:COMPILING THE C++ CODE FOR MODEL anon_model_0adcb518bef5d791c7f017c8477609d9 NOW.\n", "/anaconda3/envs/introBayesReg/lib/python3.6/site-packages/Cython/Compiler/Main.py:367: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /var/folders/rs/r1721d553r9_04l5k_315d6m0000gn/T/tmpc07gx498/stanfit4anon_model_0adcb518bef5d791c7f017c8477609d9_2519964870127563491.pyx\n", " tree = Parsing.p_module(s, pxd, full_module_name)\n" ] } ], "source": [ "boston_code = \"\"\"\n", "data {\n", " int N; // number of observations\n", " vector[N] y; // response variable\n", " int K; // number of columns in the design matrix X\n", " matrix [N, K] X; // X should not have intercept\n", " real scale_beta0; // hyperparameter(scale) on beta0\n", " vector[K] scale_betas; // hyperparameters(scale) on other betas\n", " real loc_sigma; // hyperparameter(location) on sigma\n", "}\n", "parameters {\n", " // regression coefficient vector\n", " real beta0;\n", " vector[K] betas;\n", " real sigma;\n", "}\n", "transformed parameters {\n", " vector[N] mu;\n", " mu = beta0 + X * betas;\n", "}\n", "model {\n", " // priors\n", " beta0 ~ normal(0., scale_beta0);\n", " betas ~ normal(50., scale_betas);\n", " sigma ~ exponential(loc_sigma);\n", " y ~ normal(mu, sigma); // likelihood\n", "}\n", "\"\"\"\n", "# list of predictor variables\n", "x = ['age', 'lstat']\n", "boston_dat = {'N': design_mat[x].shape[0],\n", " 'K': 2,\n", " 'y': boston_df['medv'].values,\n", " 'X': design_mat[x].values,\n", " 'scale_beta0': 100,\n", " 'scale_betas': np.array([1, 1]),\n", " 'loc_sigma': 1\n", "}\n", "#\n", "boston_sm = pystan.StanModel(model_code=boston_code)\n", "#\n", "#\n", "boston_bayesfit = boston_sm.sampling(data=boston_dat,\n", " iter=2000,\n", " chains=4,\n", " warmup=500,\n", " thin=5,\n", " seed=2019)\n", "# to summarize results\n", "fit_df = boston_bayesfit.to_dataframe()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Once you have the large sampled dataset, you summarize the results and interpret them" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'value')" ] }, "execution_count": 12, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Traceplots - simply draw index with values\n", "plt.plot(fit_df['betas[1]'])\n", "plt.title('trace plot - age', fontsize=18)\n", "plt.xlabel('sample index', fontsize=18)\n", "plt.ylabel('value', fontsize=18)" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Text(0, 0.5, 'Density')" ] }, "execution_count": 13, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Also draw density plot for samples\n", "sns.kdeplot(fit_df['betas[1]'], legend=False)\n", "plt.title('Density plot - age', fontsize=18)\n", "plt.xlabel('Value', fontsize=18)\n", "plt.ylabel('Density', fontsize=18)" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [], "source": [ "# posterior parameter distributions\n", "summary_dict = boston_bayesfit.summary()\n", "df = pd.DataFrame(summary_dict['summary'], \n", " columns=summary_dict['summary_colnames'], \n", " index=summary_dict['summary_rownames'])\n", "beta0_mean = df['mean']['beta0']\n", "beta_age_mean = df['mean']['betas[1]']\n", "beta_lstat_mean = df['mean']['betas[2]']\n", "beta0_025 = df['2.5%']['beta0']\n", "beta0_975 = df['97.5%']['beta0']\n", "beta_age_025 = df['2.5%']['betas[1]']\n", "beta_age_975 = df['97.5%']['betas[1]']\n", "beta_lstat_025 = df['2.5%']['betas[2]']\n", "beta_lstat_975 = df['97.5%']['betas[2]']" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "beta_age: mean=0.0234 || 2.5% CI=0.0015 || 97.5% CI=0.0458\n", "beta_lstat: mean=-0.9280 || 2.5% CI=-1.0197 || 97.5% CI=-0.8315\n", "intercept: mean=32.6651 || 2.5% CI=31.2583 || 97.5% CI=34.0150\n" ] } ], "source": [ "# print out results\n", "print(\"beta_age: mean={:.4f} || 2.5% CI={:.4f} || 97.5% CI={:.4f}\".format(\n", " beta_age_mean, beta_age_025, beta_age_975))\n", "print(\"beta_lstat: mean={:.4f} || 2.5% CI={:.4f} || 97.5% CI={:.4f}\".format(\n", " beta_lstat_mean, beta_lstat_025, beta_lstat_975))\n", "print(\"intercept: mean={:.4f} || 2.5% CI={:.4f} || 97.5% CI={:.4f}\".format(\n", " beta0_mean, beta0_025, beta0_975))" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "y_pred_mean is: 24.3209\n", "y_pred_sigma is 6.1734\n" ] }, { "data": { "text/plain": [ "" ] }, "execution_count": 16, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# We need to provide with X values\n", "# Posterior predictive distribution: Idea 1 - Use means of parameter samples\n", "AGE = 40\n", "LSTAT = 10\n", "y_pred_mean1 = np.mean(fit_df['betas[1]']) * AGE + np.mean(fit_df['betas[2]']) * LSTAT + np.mean(fit_df['beta0'])\n", "y_pred_sigma1 = np.mean(fit_df['sigma'])\n", "print(\"y_pred_mean is: {:.4f}\".format(y_pred_mean1))\n", "print(\"y_pred_sigma is {:.4f}\".format(y_pred_sigma1))\n", "y_pred1 = np.random.normal(y_pred_mean1, y_pred_sigma1, size=1200)\n", "plt.title(\"Idea 1. Y for age={:.0f}, lstat={:.0f}\".format(AGE, LSTAT), fontsize=18)\n", "sns.kdeplot(y_pred1)" ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "sample size: 1200\n", "Mean of posterior Y for given data: 23.9408\n", "Variance of posterior Y: 39.0989\n" ] }, { "data": { "text/plain": [ "" ] }, "execution_count": 18, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "# Posterior predictive distribution: Idea 2 - Sample Y from each parameter sample\n", "y_pred_mean2 = fit_df['betas[1]'] * AGE + fit_df['betas[2]'] * LSTAT + fit_df['beta0']\n", "y_pred_sigma2 = fit_df['sigma']\n", "y_pred2 = np.random.normal(y_pred_mean2, y_pred_sigma2)\n", "print(\"sample size:\", len(y_pred1))\n", "print(\"Mean of posterior Y for given data: {:.4f}\".format(np.mean(y_pred2)))\n", "print(\"Variance of posterior Y: {:.4f}\".format(np.var(y_pred2)))\n", "plt.title(\"Idea 2. Y for age={:.0f}, lstat={:.0f}\".format(AGE, LSTAT), fontsize=18)\n", "sns.kdeplot(y_pred2)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "So the second distribution has larger variance, because it accounts for variability of betas and sigma." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Data for Logistic Regression" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
xyposagenetusetreatedgreenphc
0349631.31458055117830040.851
1349631.3145805504041040.851
2349631.3145805504521040.851
3349631.3145805515661040.851
4349631.3145805505981040.851
\n", "
" ], "text/plain": [ " x y pos age netuse treated green phc\n", "0 349631.3 1458055 1 1783 0 0 40.85 1\n", "1 349631.3 1458055 0 404 1 0 40.85 1\n", "2 349631.3 1458055 0 452 1 0 40.85 1\n", "3 349631.3 1458055 1 566 1 0 40.85 1\n", "4 349631.3 1458055 0 598 1 0 40.85 1" ] }, "execution_count": 19, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Can use the following for logistic regression\n", "gambia_df = pd.read_csv(\"gambia.csv\") \n", "gambia_df.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Reference:\n", "- Boston dataset (R package MASS): https://cran.r-project.org/web/packages/MASS\n", "- gambia dataset (R package geoR): https://cran.r-project.org/web/packages/geoR/index.html\n", "- Stan documentation: https://mc-stan.org/users/documentation/\n", "- PyStan documentation: https://buildmedia.readthedocs.org/media/pdf/pystan/latest/pystan.pdf\n", "- PyStan linear regression: https://jrnold.github.io/bayesian_notes/introduction-to-stan-and-linear-regression.html" ] }, { "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": "3.6.7" } }, "nbformat": 4, "nbformat_minor": 2 }