Speaker:
Irene van Woerden
Title:
MCMC on Contingency Tables to find P-values
Abstract:
By running a Monte Carlo Markov Chain (MCMC) repeatedly on an observed table 
we can calculate how likely the observed table is to occur under the 
independent model.  I looked at the resolution of offences and the offender's 
ethnicity for drug related crimes observed in NZ in 2009 to test if resolution 
was independent of ethnicity.

Speaker:
Yang Hu
Title:
Trans-dimensional posterior samples over binomial partitions of extreme daily temperatures in South Island NZ
Abstract:
This report applies a new trans-dimensional extension of the rejection sampler 
of Von Neumann which provides a universal method to obtain exact samples from 
a large class of trans-dimensional target densities.  In particular, I am 
interested in applying this sampler to draw trans-dimensional posterior 
samples over binomial partitions to examine the extreme temperature events in 
five different weather stations in South Island NZ.

Speaker:
Fiona Morrison
Title:
Phylogenetic MCMC Algorithms in MrBayes
Abstract:
This project looks at convergence of MCMC methods and explores the hypothesis 
that they take an exponential number of iterations to converge under certain 
circumstances on mixtures of trees. Both the Jukes Cantor and Kimura's 
two-parameter mutation models are considered and extended TBR is used to 
change the tree topologies. Some possible solutions are also suggested.

Speaker: 
Nur Azhari
Title:
Posterior Probability for human, chimpanzee and gorilla in phylogenetic binary rooted trees
Abstract:
Reconstruction of phylogenies aid in inferring the evolutionary relationships 
among present-day organisms.  Here we study the ancestral relationship among 
human, chimpanzee and gorilla based on their DNA sequences at 782 loci from 
X chromosome.  Toward this we obtain ten million exact samples from the 
posterior distribution over the rooted binary tree spaces with three leaves.

Speaker: 
Jennifer Harlow
Title: 
Sequential adaptive approximate Bayesian computation in less than 10 minutes
Abstract: 
An outline of the sequential adaptive approximate Bayesian computation 
algorithm for making inferences about population history in a population 
genetics context.