Bayesian Non-parametric and Kingman's Coalescent
Consider a population of individuals of various types. The type frequencies evolve according to the Fleming-Viot process with parent independent mutation. Ancestral inference can be done through Kingman coalescent based on random samples taking at each fixed time. The focus of this talk is on making ancestral inferences for unobserved samples based on ancestral inference of the observed samples. This leads to the interplay between the Bayesian nonparametric analysis and Kingman coalescent. This is a joint work with Stefano Favaro and Paul Jenkins.
Professor and Associate Chair (Statistics), Department of Mathematics and Statistics, McMaster University