Title: Poisson-Dirichlet, Dirichlet process & Ewen sampling formula approximation
Speaker: Han Gan
Speaker Info: The University of Waikato
Brief Description:
Special Note:
Abstract:
The Poisson-Dirichlet distribution is a probability measure on the infinite dimensional ordered simplex, and its applications appear in a wide variety of applications in combinatorics, population genetics and Bayesian non-parameterics, to name a few. It is also intricately related with the celebrated Ewens sampling formula. In this talk we will give error bounds in the approximation of the frequencies of types for the Wright-Fisher model with infinite alleles mutation structure. This will also lead to a bound on the error between the distribution of the partition generated from sampling from a Wright-Fisher model and the Ewens sampling formula.Date: Tuesday, May 19, 2020This talk is based on joint work with Nathan Ross (UniMelb).