ST4231 Computer Intensive Statistical Methods

Module Description

The availability of high-speed computation has led to the development of “modern” statistical methods which are implemented in the form of well-understood computer algorithms. This module introduces students to several computer intensive statistical methods and the topics include: empirical distribution and plug-in principle, general algorithm of bootstrap method, bootstrap estimates of standard deviation and bias, jack-knife method, bootstrap confidence intervals, the empirical likelihood for the mean and parameters defined by simple estimating function, Wilks theorem, and EL confidence intervals, missing data, EM algorithm, Markov Chain Monte Carlo methods. This module is targeted at students who are interested in Statistics and are able to meet the prerequisite.

Lecture Topics
- Review of Basic Probability and Statistics Terminology
- Random Numbers
- Generating Random Variates with Non-Uniform Distributions
- Variance Reduction Techniques
- Markov Chain Monte Carlo
- Bootstrap Methods
- Permutation Tests
- Optimization and Equation Solving

I found this module extremely interesting because it introduced us to a lot of useful algorithms. This is a compulsory core module for statistics major. If you intend to do honours, you will probably need some of the topics learnt in this module. For example, MCMC, Gibbs sampling, bootstrapping, simulation, etc


Workload
Three-hours lecture
Lectures were taught by Dr Wu Zhengxiao. The lecturers changes almost(?) every semester so I'm not sure if this is still relevant to this module. However Dr Wu is alright. We can understand him find and he teaches well.

One two-hours tutorial
Tutorials were held after the 2nd lecture. Tutorials were quite tough for my semester because there are quite a bit of coding. The questions are non-trivial and I often had to spend a lot of them on them. We used R for all the coding.

Textbook is not compulsory. Dr Wu used Computational Statistics, 2nd Edition by Geof H. Givens and Jennifer A. Hoeting.

Lectures were not webcasted.

Assessment (AY11/12 Sem 2)
20% Homework/participation/tutorial
20% Midterm
60% Finals

Dr Wu would ask someone to come out and present one question for every tutorial. So it is a good idea to do the tutorial beforehand.

1 single A4 cheat sheet is allowed for midterm. The midterm wasn't too difficult. He tested on everything in the lecture notes. I made a mistake for the question about changing variables.

2 A4 cheatsheet are allowed for finals. Final paper was significantly more difficult. Most of the questions were about writing pseudo codes.