Bayesian Methods for Regression in R

Course Topics

[video:https://vimeo.com/39411439]

LISA Short Course:Bayesian Methods for Regression in R from LISA on Vimeo.

 

An outline for questions I hope to answer:

What is Bayes’ Rule? (lecture portion)

► What is the likelihood?

► What is the prior distribution?

► How should I choose it?

► Why use a conjugate prior?

► What is a subjective versus objective prior?

► What is the posterior distribution?

► How do I use it to make statistical inference?

► How is this inference different from frequentist/classical inference?

► What computational tools do I need in order to make inference?

How can I use R to do regression in a Bayesian paradigm? (computer portion)

► What libraries in R support Bayesian analysis?

► How do I use some of these libraries?

► How do I interpret the output?

► How do I produce diagnostic plots?

► What common topics do these libraries not support?

► How can I do them myself?

► How can LISA help me?

► What resources are available to help me Bayesian methods in R?

Before you show up:

The main focus of this short course will be the Bayesian aspect of it. That means this is a slightly more advanced course requiring some knowledge of basic probability, regression methods, and the R software language. The pre course assignment is quite long. If there are already parts you are comfortable with, feel free to skip.

Pre-course assignment:

  • Refresh basics of probability
    • Conditional probability
    • Bayes’ Theorem
    • Common probability distributions
      • Normal
      • Gamma (and its special cases)
      • Poisson
      • Binomial (Bernoulli is a special case)
      • Beta
      • T
      • Uniform
    • Not so common probability distributions
      • Inverse-Gamma
      • Wishart
      • Inverse-Wishart
      • Dirichlet