Home > Academic Announcements > (Jul. 24) A new rejection sampling method without using hat function and its application in parallel big data analysis

(Jul. 24) A new rejection sampling method without using hat function and its application in parallel big data analysis

Last updated :2017-07-17

Topic: A new rejection sampling method without using hat function and its application in parallel big data analysis
Speaker: Hongsheng Dai
(University of Essex)
Time: 9:00-10:00 am, Monday, July 24, 2017
Venue: Room 415, New Mathematics Building, Guangzhou South Campus, SYSU

Abstract:
It is nontrivial to draw exact realisations from the posterior of finite mixture models, for which all existing exact Monte Carlo simulation methods are not practical or just work theoretically. Motivated by this problem, this work proposes a new exact simulation method, which simulates a realisation from a proposal density and then uses exact simulation of a Langevin diffusion to check whether the proposal should be accepted or rejected. Comparing to the existing rejection sampling method, the new method does not require the proposal density function to bound the target density function. The new method is much more efficient than existing methods for simulation from the posterior of finite mixture models. This presentation will also discuss the possibility of applying this method to Bayesian group decision theory and big data analysis via parallel computing.