MAS3301, Bayesian Statistics

Semester 2, 2008-9

MAS3301 students should "Bookmark" this page!

Reading list information at Blackwell's . The password is zigzag1dr.
Timetable
Lecture notes etc.
Assessment
Data
Information on R
Other resources

Timetable

Times and locations of classes are as follows.

===================================================================================
Day       Time      Room    Class             Weeks
===================================================================================
Tuesday   10-11      TR4   Lecture          Every week   
Wednesday 11-12      TR4   Lecture          Every week 
Thursday   3-4       TR4   Problems Class   Odd weeks 
Thursday   3-4       TR4   Drop in          Weeks 2, 6, 10, 12 
Thursday   3-4      Loch   Practical        Week 4 (Feb 19th) and Week 8 (Mar 19th)
===================================================================================

TR4: Herschel, Level 4, Teaching Room 4.
Loch: Herschel, Level 1, Loch PC Laboratory.

An attendance register will be taken at each problem or practical class. Please note that a good attendence record will be taken into account when deciding whether to pass students with borderline assignment/exam marks.

Lecture notes

Lecture notes will be posted below, together with solutions to problems. Extra lectures will be added to the notes as they are written. They are in pdf format and you will need Acrobat Reader to view them - click on the link to download your own (free) copy.

Lecture notes: Lectures 1-2 (week 1).
Solutions: Problems 1.
Lecture notes: Lectures 3-4 (week 2).
Lecture notes: Lectures 5-6 (week 3).
Solutions: Problems 2.
Lecture notes: Lectures 7-8 (week 4).
Lecture notes: Lectures 9-10 (week 5).
Solutions: Problems 3.
Lecture notes: Lectures 11-12 (week 6).
Lecture notes: Lectures 13-14 (week 7).
Solutions: Problems 4.
Lecture notes: Lectures 15-16 (week 8).
Lecture notes: Lectures 17-18 (week 9).
Solutions: Problems 5.
Lecture notes: Lectures 19-20 (week 10).
Lecture notes: Lecture 21 (week 11).

Assessment

The assessment of the module consists of a 90 minute examination at the end of Semester 2 (80%), 5 sets of assessed problems (10%) and a project (10%). Obviously I can't show you this year's exam but you can get a copy of last year's exam paper from the university's exam site or from here. The solution to B7 is here. This module was MAS354 until 2006 so you can look at earlier MAS354 papers as well.

I plan to give five pieces of assessed coursework (in addition to the project). These will be equally weighted, so each counts for 2% of the module mark. The dates will be as follows. Work must be handed in by 4.00 pm on the dates given below, in the homework letterbox at the Mathematics and Statistics General Office, 3rd floor, Herschel Building.

Assignment      Hand out      Hand in 
---------------------------------------------------------
    1            Week 1       Week 3 (Monday February 9) 
    2            Week 3       Week 5 (Monday February 23) 
    3            Week 5       Week 7 (Monday March 9) 
    4            Week 7       Week 9 (Monday April 20) 
    5            Week 9       Week 11 (TUESDAY May 5)

The Project is to be submitted by 4.00pm on Thursday 30th April (Week 10). It must be submitted to the General Office and you will need to sign a sheet when you submit it. I will give you the project by the end of Week 3. Please note that in Weeks 9, 10, 11 you will have three pieces of work to submit, two assignments and the project, so try to make a start on the project well before the hand-in date. You will have 7 weeks of term and 4 weeks of the Easter vacation to work on the project. That should be more than enough.

The project instruction may be obtained here.
Here is the outline solution to this year's project.

Data

Chester Road traffic data.
Weibull example data (weibex.txt).
Cholesterol data (cholesterol.txt).

Information on R

An Introduction to R can be downloaded here.

R function for Weibull example: weibpost.
R function for probit example: probit1.
R function for probit example: probit2.
R function for gamma hpd interval: hpdgamma.
R functions for finding gamma distribution parameters when two quantiles are given: findab, findb.
R function for beta hpd interval: hpdbeta.
R function for several normal samples, conjugate prior: oneway.

Other resources

A series of talks on Bayesian methods in health economics can be found here here. They may be of interest.

Please report to me any errors or suggested improvements to these notes.