1st Edition

# Probability and Bayesian Modeling

By Jim Albert, Jingchen Hu Copyright 2020
552 Pages
by Chapman & Hall

552 Pages
by Chapman & Hall

Also available as eBook on:

Probability and Bayesian Modeling is an introduction to probability and Bayesian thinking for undergraduate students with a calculus background. The first part of the book provides a broad view of probability including foundations, conditional probability, discrete and continuous distributions, and joint distributions. Statistical inference is presented completely from a Bayesian perspective. The text introduces inference and prediction for a single proportion and a single mean from Normal sampling. After fundamentals of Markov Chain Monte Carlo algorithms are introduced, Bayesian inference is described for hierarchical and regression models including logistic regression. The book presents several case studies motivated by some historical Bayesian studies and the authors’ research.

This text reflects modern Bayesian statistical practice. Simulation is introduced in all the probability chapters and extensively used in the Bayesian material to simulate from the posterior and predictive distributions. One chapter describes the basic tenets of Metropolis and Gibbs sampling algorithms; however several chapters introduce the fundamentals of Bayesian inference for conjugate priors to deepen understanding. Strategies for constructing prior distributions are described in situations when one has substantial prior information and for cases where one has weak prior knowledge. One chapter introduces hierarchical Bayesian modeling as a practical way of combining data from different groups. There is an extensive discussion of Bayesian regression models including the construction of informative priors, inference about functions of the parameters of interest, prediction, and model selection.

The text uses JAGS (Just Another Gibbs Sampler) as a general-purpose computational method for simulating from posterior distributions for a variety of Bayesian models. An R package ProbBayes is available containing all of the book datasets and special functions for illustrating concepts from the book.

A complete solutions manual is available for instructors who adopt the book in the Additional Resources section.

1. Probability: A Measurement of Uncertainty
2. Introduction

The Classical View of a Probability

The Frequency View of a Probability

The Subjective View of a Probability

The Sample Space

Assigning Probabilities

Events and Event Operations

The Three Probability Axioms

Exercises

3. Counting Methods
4. Introduction: Rolling Dice, Yahtzee, and Roulette

Equally Likely Outcomes

The Multiplication Counting Rule

Permutations

Combinations

Arrangements of Non-Distinct Objects

Playing Yahtzee

Exercises

5. Conditional Probability
6. Introduction: The Three Card Problem

In Everyday Life

In a Two-Way Table

Definition and the Multiplication Rule

The Multiplication Rule Under Independence

Learning Using Bayes' Rule

R Example: Learning About a Spinner

Exercises

7. Discrete Distributions
8. Introduction: The Hat Check Problem

Random Variable and Probability Distribution

Summarizing a Probability Distribution

Standard Deviation of a Probability Distribution

Coin-Tossing Distributions

Binomial probabilities

Binomial computations

Mean and standard deviation of a Binomial

Negative Binomial Experiments

Exercises

9. Continuous Distributions
10. Introduction: A Baseball Spinner Game

The Uniform Distribution

Probability Density: Waiting for a Bus

The Cumulative Distribution Function

Summarizing a Continuous Random Variable

Normal Distribution

Binomial Probabilities and the Normal Curve

Sampling Distribution of the Mean

Exercises

11. Joint Probability Distributions
12. Introduction

Joint Probability Mass Function: Sampling From a Box

Multinomial Experiments

Joint Density Functions

Independence and Measuring Association

Flipping a Random Coin: The Beta-Binomial Distribution

Bivariate Normal Distribution

Exercises

13. Learning About a Binomial Probability
14. Introduction: Thinking About a Proportion Subjectively

Bayesian Inference with Discrete Priors

Example: students' dining preference

Discrete prior distributions for proportion p

Likelihood of proportion p

Posterior distribution for proportion p

Inference: students' dining preference

Discussion: using a discrete prior

Continuous Priors

The Beta distribution and probabilities

Choosing a Beta density curve to represent prior opinion

Updating the Beta Prior

Bayes' rule calculation

From Beta prior to Beta posterior: conjugate priors

Bayesian Inferences with Continuous Priors

Bayesian hypothesis testing

Bayesian credible intervals

Bayesian prediction

Predictive Checking

Exercises

15. Modeling Measurement and Count Data
16. Introduction

Modeling Measurements

Examples

The general approach

Outline of chapter

Bayesian Inference with Discrete Priors

Example: Roger Federer's time-to-serve

Simplification of the likelihood

Inference: Federer's time-to-serve

Continuous Priors

The Normal prior for mean _

Choosing a Normal prior

Updating the Normal Prior

Introduction

A quick peak at the update procedure

Bayes' rule calculation

Conjugate Normal prior

Bayesian Inferences for Continuous Normal Mean

Bayesian hypothesis testing and credible interval

Bayesian prediction

Posterior Predictive Checking

Modeling Count Data

Examples

The Poisson distribution

Bayesian inferences

Case study: Learning about website counts

Exercises

17. Simulation by Markov Chain Monte Carlo
18. Introduction

The Bayesian computation problem

Choosing a prior

The two-parameter Normal problem

Overview of the chapter

Markov Chains

Definition

Some properties

Simulating a Markov chain

The Metropolis Algorithm

Example: Walking on a number line

The general algorithm

A general function for the Metropolis algorithm

Example: Cauchy-Normal problem

Choice of starting value and proposal region

Collecting the simulated draws

Gibbs Sampling

Bivariate discrete distribution

Beta-Binomial sampling

Normal sampling { both parameters unknown

MCMC Inputs and Diagnostics

Burn-in, starting values, and multiple chains

Diagnostics

Graphs and summaries

Using JAGS

Normal sampling model

Multiple chains

Posterior predictive checking

Comparing two proportions

Exercises

19. Bayesian Hierarchical Modeling
20. Introduction

Observations in groups

Example: standardized test scores

Separate estimates?

Combined estimates?

A two-stage prior leading to compromise estimates

Hierarchical Normal Modeling

Example: ratings of animation movies

A hierarchical Normal model with random _

Inference through MCMC

Hierarchical Beta-Binomial Modeling

Example: Deaths after heart attack

A hierarchical Beta-Binomial model

Inference through MCMC

Exercises

21. Simple Linear Regression
22. Introduction

Example: Prices and Areas of House Sales

A Simple Linear Regression Model

A Weakly Informative Prior

Posterior Analysis

Inference through MCMC

Bayesian Inferences with Simple Linear Regression

Simulate fits from the regression model

Prediction of future response

Posterior predictive model checking

Informative Prior

Standardization

Prior distributions

Posterior Analysis

A Conditional Means Prior

Exercises

23. Bayesian Multiple Regression and Logistic Models
24. Introduction

Bayesian Multiple Linear Regression

Example: expenditures of US households

A multiple linear regression model

Weakly informative priors and inference through MCMC

Prediction

Comparing Regression Models

Bayesian Logistic Regression

Example: US women labor participation

A logistic regression model

Conditional means priors and inference through MCMC

Prediction

Exercises

25. Case Studies

Introduction

Federalist Papers Study

Introduction

Data on word use

Poisson density sampling

Negative Binomial sampling

Comparison of rates for two authors

Which words distinguish the two authors?

Career Trajectories

Introduction

Measuring hitting performance in baseball

A hitter's career trajectory

Estimating a single trajectory

Estimating many trajectories by a hierarchical model

Latent Class Modeling

Two classes of test takers

A latent class model with two classes

Disputed authorship of the Federalist Papers

Exercises

Appendix

Appendix A: The constant in the Beta posterior

Appendix B: The posterior predictive distribution

Appendix C: Comparing Bayesian models

### Biography

Jim Albert is a Distinguished University Professor of Statistics at Bowling Green State University. His research interests include Bayesian modeling and applications of statistical thinking in sports. He has authored or coauthored several books including Ordinal Data Modeling, Bayesian Computation with R, and Workshop Statistics: Discovery with Data, A Bayesian Approach.

Jingchen (Monika) Hu is an Assistant Professor of Mathematics and Statistics at Vassar College. She teaches an undergraduate-level Bayesian Statistics course at Vassar, which is shared online across several liberal arts colleges. Her research focuses on dealing with data privacy issues by releasing synthetic data.

"The book can be used by upper undergraduate and graduate students as well as researchers and practitioners in statistics and data science from all disciplines…A background of calculus is required for the reader but no experience in programming is needed. The writing style of the book is extremely reader friendly. It provides numerous illustrative examples, valuable resources, a rich collection of materials, and a memorable learning experience."
~Technometrics

"Over many years, I have wondered about the following: Should a first undergraduate course in statistics be a Bayesian course? After reading this book, I have come to the conclusion that the answer is…yes!... this is very well written textbook that can also be used as self-learning material for practitioners. It presents a clear, accessible, and entertaining account of the interplay of probability, computations, and statistical inference from the Bayesian perspective."
~ISCB News