用到的书是Bayesian Data Analysis Third Edition

This course introduces the Bayesian paradigm for statistical inference. Topics covered include prior and posterior distributions: conjugate priors, informative and non-informative priors; one- and two-sample problems; models for normal data, models for binary data, Bayesian linear models, Bayesian computation: MCMC algorithms, the Gibbs sampler; hierarchical models; hypothesis testing, Bayes factors, model selection; use of statistical software.

Prerequisites: A course in the theory of statistical inference, such as STAT GU4204/GR5204 a course in statistical modeling and data analysis such as STAT GU4205/GR5205.

对Bayes有初步了解的，有一定 statistics，machine learning 基础的，且对理论有兴趣的人。整体上是基于 measure-theoretical probability， 所以读者最好有一定 measure theory基础，基本上，具有数学本科三年级基础就足够了。主要内容:
Parametric Bayes and posterior consistency.
Non-parametric Bayes, including Dirichlet process prior and Gaussian process prior.
General consistency theory for non-parametric Bayes.
Reproducing kernel Hilbert space (RKHS) theory.

5 Hierarchical models

5.1 Constructing a parameterized prior distribution
5.2 Exchangeability and setting up hierarchical models
5.3 Fully Bayesian analysis of conjugate hierarchical models
5.4 Estimating exchangeable parameters from a normal model
5.5 Example: parallel experiments in eight schools
5.6 Hierarchical modeling applied to a meta-analysis
5.7 Weakly informative priors for hierarchical variance parameters
5.8 Bibliographic note
5.9 Exercises

6 Model checking

6.1 The place of model checking in applied Bayesian statistics
6.2 Do the inferences from the model make sense?
6.3 Posterior predictive checking
6.4 Graphical posterior predictive checks
6.5 Model checking for the educational testing example
6.6 Bibliographic note