vamads/STT465

STT 465 : Bayesian Statistical Methods (MSU)

STT 465: Bayesian Statistical Methods, Fall 2017

Instructor: Gustavo de los Campos ( EPI-BST, STT & IQ ).

Contact: gustavoc@msu.edu  +1(517)884-7607

Office: 775 Woodlot Dr. Office 1311, MI 48824

Time/Place: MWF 10:20am - 11:10am A222 Wells Hall

Course Website: https://github.com/gdlc/STT465

Office Hours: I would like to discuss all the questions that you have within the class. I will be sure to provide plenty of time during class to discuss any questions that you may have. Additionally I'll be available for office hours Wednesday & Friday 9:00-10:00am (@ Wells Hall room 404C).

Textbook (required): A First Course in Bayesian Statistical Methods, Peter D. Hoff, Springer.

Syllabus

Tentative Schedule

Topics

• Statistical Inference (Why? and How?)

• Review of Probability & Bayes Theorem

• Introduction to R [Covered by Marco Lopez-Cruz].

• Beta-Binomial Model

• Poisson-Gamma Model

• Normal Model

  • Inferering the mean, assuming known variance
  • Infering both the mean and the variance via Gibbs Sampling

• Multiple Linear Regression

  • Handout (includes OLS, Max. Likelihood & Bayesian).
  • Gibbs Sampler
  • Post-Gibbs
  • EXAMPLE

• Logistic Regression

  • Handout
  • R-examples
  • Metropolis algorithm

• Finite Mixture Models

Homework

• HW 1 (probability) (Due Sunday Sept. 16th).

• HW 2 (Binomial & Poisson model) (Due Sunday Sept. 30th).

• HW 3 (Poisson & Normal models) (Due Wed. Oct. 9th).

• HW 4 (Linear regression OLS).

• HW 5 (Linear regression Bayesian).

• HW 6 (Logistic regression, likelihood & Bayesian). (Due Monday Dec. 3rd)

• Sample Answers: HW1, HW2 and HW3