Thesis and related work

CARME 2019 presentation

SASA2018 presentation

The main result of my thesis is best explained by this poster.

The work primarily consists of 5 papers, linked below.

  1. Bayesian Fitting of Dirichlet Type I and II Distributions
  2. Multiple Discrimination Between Dirichlet Populations
  3. Classification of Multiple Dirichlet Observations under a Multinomial Model
  4. A Multivariate Gamma Distribution Applied To Composition Data Analysis
  5. A method for Bayesian regression modelling of composition data

Firstly we have a post with generally useful code for the Dirichlet distribution.

Then we have a document with commented code explaining the example on Netball data in the 5th paper above (Chapter 6 of the thesis), plus its accompanying dataset.


Those interested in writing a thesis in LateX might like to have a look at my source for the preamble and introduction.

The bibliograph file might be useful to somebody doing research in this field.

Sean van der Merwe
Coordinator of UFS Statistical Consultation Unit