Computational Bayesian Inference

The techniques of statistical inference for studying properties of data generating processes include method of moments, maximum likelihood estimation, Bayesian inference, and nonparametric statistics.

Bayesian inference is an important approach to data analysis in which unknown parameters are treated as random variables whose probability distributions can be updated in light of new information.

Bayesian inference is particularly advantageous for sequential data analysis and hypothesis testing when data are being collected sequentially.

In this course, students learn foundations of Bayesian inference, including contemporary computational methods such as Markov Chain Monte Carlo (MCMC) and get hands-on experience using R.

Topics covered in the course include Bayes' rule, prior and posterior distributions, Markov Chain (MC), MCMC methods, the celebrated Metropolis-Hastings algorithm, and the Gibbs sampler.

Students may not take both CSCI E-161 and ISMT E-161 (offered previously) for degree or certificate credit.