American Statistical Association
My talk aims to be a how-to guide on implementing the Dirichlet Process (DP) from scratch. I plan to summarize the essential ingredients in implementing DP: 1) the proof of exchangeability that allows DP to spawn new clusters as data accrue; and 2) the MCMC algorithm using the Chinese Restaurant Process. These essential ingredients are often omitted, not clearly explained, or simply assumed understood in tutorials you find online. I will walk you through a working example step by step. Important derivations missing in popular tutorials will be restored. The goal is to give you a good starting point in incorporating Bayesian nonparametrics in your own methods development.
|Date:||Wednesday, October 10, 2018|
|Time:||4:00 - 5:00 P.M.|
Memorial Sloan Kettering Cancer Center
Department of Epidemiology and Biostatistics
485 Lexington Avenue
(Between 46th & 47th Streets)
2nd Floor, Conference Room B
New York, New York
**Outside visitors please email email@example.com for building access.
You must be on the security list to enter the floor.