{"success":1,"msg":"","color":"rgb(28, 35, 49)","title":"Adaptive Random Neighbourhood MCMC schemes for large variable selection problems<\/b>","description":"webinar","title2":"","start":"2021-10-15 13:00","end":"2021-10-15 14:00","responsable":"Vladimir Rodriguez-Caballero <\/i><\/a>","speaker":"Jim Griffin. UCL","id":"53","type":"webinar","timezone":"America\/Mexico_City","activity":"Join Zoom Meeting\r\nhttps:\/\/itam.zoom.us\/j\/97039403155?pwd=UG90ajRKYjdiRWd5eldZdFowS0gwQT09\r\n\r\nMeeting ID: 970 3940 3155\r\nPasscode: 886009","abstract":"Data set with many variables (often, in the hundreds, thousands, or more) are routinely collected in many disciplines. This has motivated the study of variable selection in regression models with a large number of variables. A standard Bayesian approach defines a prior on the model space and uses Markov chain Monte Carlo methods to sample the posterior. Unfortunately, simple, default Markov chain Monte Carlo methods often mix poorly. \r\n\r\nIn this talk, I will describe several adaptive Metropolis-Hastings schemes built around the idea of proposing from a random neighborhood around the current model. I will discuss the ability of these methods to sample from the posterior in high-dimensional problems. The methods will be illustrated on simulated and real data with hundreds or thousands of variables.\r\n"}