Dr Konstantinos Kalogeropoulos COL.6.10

This course is available on the BSc in Accounting and Finance, BSc in Actuarial Science, BSc in Business Mathematics and Statistics, BSc in Mathematics with Economics and BSc in Statistics with Finance. This course is available as an outside option to students on other programmes where regulations permit and to General Course students.

Students must have completed Mathematical Methods (MA100) and Elementary Statistical Theory (ST102).

ST202 is also recommended.

Statistical decision theory: risk, decision rules, loss and utility functions, Bayesian expected loss, Frequentist risk.

Bayesian Analysis: Bayes theorem, prior, posterior and predictive distributions, conjugate models (Normal-Normal, Poisson-Gamma, Beta-Binomial), Bayesian point estimation, credible intervals and hypothesis testing, Bayes factors and model selection. Comparison with Frequentist approaches.

Implementation: Asymptotic approximations (Laplace approximation, Monte Carlo methods, stochastic simulation), Markov Chain Monte Carlo (MCMC) simulation (Gibbs sampler, Metropolis-Hastings algorithm). Computer tools (R, WinBUGS). Illustration via applications in Regression (Linear, ANOVA, Multiple, Generalized Linear Models), Hierarchical/ Multilevel Models and Time Series.

20 hours of lectures and 10 hours of computer workshops in the LT. 1 hour of lectures in the ST.

There will be a reading week in week 6.

Optional problem sets and computer exercises.

P.M. Lee, Bayesian Statistics. An introduction.

J.O. Berger, Statistical Decision Theory and Bayesian Analysis.

D. Gamerman, H. F. Lopes, Markov Chain Monte Carlo: Stochastic Simulation for Bayesian Inference

A. Gelman, Bayesian data analysis.

Exam (80%, duration: 2 hours) in the main exam period.
Project (20%) in the ST.