# STATISTICS 225 - Bayesian Statistical Analysis

## TuTh 12:30pm - 1:50pm in DBH 1423

```
instructor: Hal Stern
office: 2216 Bren Hall
phone: 949-824-1568
email: sternh@uci.edu
office hours: M 3:00-4:00pm

Course documents will be made available here:

Syllabus
Computing handout

Examples from class (including R code):

One parameter example in R
Logistic regression (2-parameter example) in R (grid sampling, Stan)
Stan code for the logistic regression example
Normal-normal hierarchical model (8 schools) in R (grid sampling, Stan)
Stan code for the normal-normal hierarchical model
Beta-binomial hierarchical model (70 rat studies) in R (grid sampling, Stan)
Stan code for beta-binomial w/ default prior
Stan code for beta-binomial w/ informative prior
R code to compute Gelman & Rubin convergence diagnostics
Gibbs sampling for normal-normal model in R
Finding marginal posterior mode for normal-normal model in R (used to identify starting points for Metropolis algorithm)
Metropolis algorithm for normal-normal model in R
Variational Bayes output displays (from text)
Hierarchical regression example

Homework assignments:

Homework 1 (assigned 10/4, due 10/16)
Solutions and R code/ouput (posted 10/25)
Homework 2 (assigned 10/18, due 10/30)
Solutions and R code/output (posted 11/4)
Homework 3 (assigned 11/1, due 11/15)
Note: Problem 1 of HW 3 is due 11/8 (it concerns project proposals)
Solutions and code (Metropolis, Gibbs)
Homework 4 (assigned 11/15, due 11/29)
Solutions and code

Exam information:

Old exam
Solutions to old exam

Exam
Solutions to exam

Lecture notes:

These files are sets of slides that serve
as "approximate" lecture notes.  Each set should last about
two-three weeks.

Slides #1 - Introduction, univariate models, multivariate models, large sample results
Slides #2 - Hierarchical models, Bayesian computing, model checking, classical vs Bayes
Slides #3 - Robust models, regression models (regular and hierarchical), data collection