STAT 4010 - Bayesian Learning
2024 Spring
Class Information
Class time: Th 0930-1215
Location: ERB 407
E-learning platform: http://www1.sta.cuhk.edu.hk/
Instructor
Name: Kin Wai (Keith) CHAN
Email: kinwaichan@cuhk.edu.hk
Office: LSB 115
Tel: 3943 7923
Office hour: Open-door policy
Teaching Assistants
Office: LSB G32
Tel: 3943 8535
Email: 1155125610@link.cuhk.edu.hk
Office: LSB 130
Tel: 3943 7939
Description
This course introduces the basic Bayesian inference procedures and philosophy, emphasizing both conceptual foundations and implementation. It covers conjugate families of distributions, Bayesian credible region, Jeffery’s prior, Markov Chain Monte Carlo, Bayes factor, Bayesian information criterion, imputation, Bayesian linear-regression models, model averaging, hierarchical models and empirical Bayes models. Hands-on implementation of estimation and inference procedures in R will be demonstrated in interactive sections.
Prerequisites
No prerequisite course, but probability, statistics, and programming knowledge at the level of Stat 2001, 2005, and 2006 is highly recommended.
Textbooks
A self-contained lecture note is the main source of reference. Complementary textbooks include
(Major) Hoff, P.D. (2009). A First Course in Bayesian Statistical Methods. Springer. (Free online access via CUHK library)
(Minor) Wasserman, L. (2004). All of Statistics: A Concise Course in Statistical Inference. Springer.
(Minor) Albert, J. (2007). Bayesian Computation with R. Springer.
Learning outcomes
Upon finishing the course, students are expected to
distinguish the difference between frequentist and Bayesian methods, and identify their pros and cons;
derive posterior distribution from commonly-used prior and sampling distributions;
perform Markov chain Monte Carlo for Bayesian inference in R;
build simple Bayesian models for solving real problems; and
select appropriate Bayesian tools for different statistical tasks, e.g., estimation, testing, model selection, prediction, etc.
Assessment and Grading
There are three main assessment components, plus a bonus component.
a (out of 100) is the average score of approximately eight assignments with the lowest two scores dropped;
m (out of 100) is the score of mid-term assessment;
f (out of 100) is the score of final assessment; and
b (out of 2) is the bonus points, which will be given to students who actively participate in class.
The total score t (out of 100) is given by
t = min{100, 0.3a + 0.2max(m,f ) + 0.5f + b}
If min(t, f ) < 30, the final letter grade will be handled on a case-by-case basis. Otherwise, your letter grade will be in the A range if t ≥ 85, at least in the B range if t ≥ 65, at least in the C range if t ≥ 55.
Important note: For the most updated information, please always refer to the course outline announced by the course instructor in Blackboard, which shall prevail the above information if there is any discrepancy.
Lecture Notes
All rights reserved by the authors. Re-distribution by any means is strictly prohibited
Front matters
Click (S4010/lecture) to download all lecture notes and codes (or click the individual links below).
Part 1: Basics of Bayesian Inference
Part 2: Theory and Computation
Part 3: Applications
Appendixes (Optional)
Appendix A: Basic Mathematics (Optional) --- for students who want to review; read Chapter A in STAT 4003
Appendix B: Basic probability and statistics (Optional) --- for students who want to review; read Chapters 1 and 2 in STAT 4003
Appendix C: Basic Programming in R (Optional) --- for students who want to review; read Lectures 2 and 3 in RMSC 1101
Appendix D: Why Isn't Everyone a Bayesian? --- Efron (1986) (Optional)
Special topics that may be covered: Bayesian nonparametric, Fiducial inference, convergence diagnosis on MCMC, ...
In-class Materials
Click (S4010/2024Spring/L/inclassNote) and (S4010/2024Spring/L/recording) to download all in-class notes and recordings (if any).
Lecture 1 (11 Jan): Introduction to Bayesian Statistics, posterior calculation. [Sorry that the screen was not recorded this time]
Lecture 2 (18 Jan): Examples of posterior calculation, introduction to non-informative prior. [Example 1.4 of the lecture note was updated.]
Lecture 3 (25 Jan): Invariant priors, Jeffreys priors, violation of likelihood principle, conjugate priors and informative priors.
Lecture 4 (1 Feb): Conjugate prior for exponential family, mixture distribution, commonly used Bayesian estimators, decision theory.
Lecture 5 (8 Feb): Derivation of commonly used Bayes Estimators, admissibility (Mini-talk of luck)
Lecture 6 (22 Feb): Admissibility, Minimaxity, proofs of Chapter 3. [Sorry that the sound was not recorded this time]
Lecture 7 (29 Feb): 0-1 loss for testing, generalized 0-1 loss, Bayes factor, modified Bayes factor
Lecture 8 (14 Mar): Problems of Bayesian testing, region estimation, HPD credible region
Lecture 9 (21 Mar): Asymptotic theory, De Finetti's theorem, classical and MC methods
Lecture 10 (28 Mar): Computational methods, Example 7.8
Lecture 11 (11 Apr): Convergence diagnosis, Gibbs sampler, MH-within-Gibbs sampler
Lecture 12 (18 Apr): Regression, BIC, spike-and-slab prior, LASSO
Lecture 13 (22 Apr): Missing data, EM, MI, empirical Bayes
Remark: In-class notes and recordings (if any) will be uploaded within one week after the lecture.
Are you a Bayesian or frequentist?
11 Jan 2024: B vs F = 65% vs 35%
Assignments
Click (S4010/2024Spring/A) to download all assignments.
Assignment 1: posterior calculation, representation --- Due: 2 Feb (Fri) @ 1800
Assignmnet 2: prior specification, other types of priors --- Due: 16 Feb (Fri) @ 1800
Assignment 3: Bayesian estimators --- Due: 1 Mar (Fri) @ 1800
Assignment 4: Bayesian testing and region estimation with application in horse racing --- Due: 29 Mar (Fri) @ 1800
Assignment 5: Theoretic comparison between Bayesian and frequentist's methods --- Due: 5 Apr (Fri) @ 1800
Assignment 6: MCMC and parallel computing --- Due: 26 April (Fri) @ 1800
Assignment 7: MCMC and Probit regression --- Due: 30 Apr (Tue) @1800
Quizzes
Click (S4010/2024Spring/Q) to download all assignments.
Quiz 1: Posterior calculation
Quiz 2: Prior specification
Quiz 3: Bayes estimation
Mid-term Project
Start time: 1 March (Friday), 7:00 pm
End time: 3 March (Sunday), 7:00 pm
Duration: 48 hours
Scope: Chapters 1-3
Instructions: The detailed instructions are stated on the first page of the real question paper. Some highlights are listed below:
Read the instructions carefully before doing the project.
Complete the project by yourself.
Consult and use any official course materials if you wish.
The project will be graded according to the criterion referencing scheme (see here).
Submission:
Compile your answers in a single ".pdf" file (i.e., not MS words, jpeg, zip, etc).
Sign the Honor Code, and attach it as a cover of your submitted file.
Name the document in the format S4010_M_sid_name.pdf., e.g., S4010_M_1155001234_ChanKinWai.pdf.
Submit to Blackboard. You may submit your answers as many times as you wish, however, only the last submission will be graded.
All plots, numerical answers, simulation results, etc must be included in the written part. Graders will not run your submitted codes to check the answers.
Mock mid-term projects
Mock 1: 2023 Spring M -- Most relevant
Mock 2: 2022 Spring M -- Most relevant
Mock 3: 2021 Spring M
Mock 4: 2021 Spring M0
Mock 5: 2019 Fall M
Mock 6: 2019 Fall M0
(Real) Mid-term project
Question paper -- The password will be sent to your CUHK email sharply at the project start time.
Final Project
Start time: 3 May (Friday), 10:00 am
End time: 6 May (Monday), 10:00 am
Duration: 72 hours
Scope: Chapters 1-10
Instructions: The detailed instructions are stated on the first page of the real question paper. Some highlights are listed below:
Read the instructions carefully before doing the project.
Complete the project by yourself.
Consult and use any official course materials if you wish.
The project will be graded according to the criterion referencing scheme (see here).
Submission:
Compile your answers in a single ".pdf" file (i.e., not MS words, jpeg, zip, etc).
Sign the Honor Code, and attach it as a cover of your submitted file.
Name the document in the format S4010_F_sid_name.pdf., e.g., S4010_F_1155001234_ChanKinWai.pdf.
Submit to Blackboard. You may submit your answers as many times as you wish, however, only the last submission will be graded.
All plots, numerical answers, simulation results, etc must be included in the written part. Graders will not run your submitted codes to check the answers.
Mock final projects
Quick review exercise for each chapter. (Recording from 2021 Spring)
Mock 1: 2023 Spring F -- Most relevant
Mock 2: 2022 Spring F
Mock 3: 2021 Spring F
Mock 4: 2019 Fall F
Mock 5: 2019 Fall F0
Real final project
Question paper: (S4010/2024Spring/F) -- The password will be sent to your CUHK email sharply at the project start time.