{"success":1,"msg":"","color":"rgb(28, 35, 49)","title":"<b>Adapting the Metropolis Algorithm<\/b>","description":"webinar","title2":"","start":"2021-01-29 11:00","end":"2021-01-29 12:30","responsable":"<a href=\"mailto:vladimir.rodriguez@itam.mx\">Carlos Vladimir Rodriguez-Caballero <i class=\"fa fa-envelope\"><\/i><\/a>","speaker":"Jeffrey S. Rosenthal,  University of Toronto ","id":"21","type":"webinar","timezone":"America\/Mexico_City","activity":"Zoom link:\r\nhttps:\/\/itam.zoom.us\/j\/95100639455?pwd=VE94MjRNS0FjNlJUTzhjU0x6eEJEdz09\r\n\r\nID Meeting: 951 0063 9455\r\nAccess code: 187963\r\n","abstract":"The Metropolis Algorithm is an extremely useful and popular method of\r\napproximately sampling from complicated probability distributions.\r\n\"Adaptive\" versions automatically modify the algorithm while it runs, to\r\nimprove its performance on the fly, but at the risk of destroying the\r\nMarkov chain properties necessary for the algorithm to be valid.  In this\r\ntalk, we will illustrate the Metropolis algorithm using a very simple\r\nJavaScript example (http:\/\/probability.ca\/jeff\/js\/metropolis.html).  We\r\nwill then discuss adaptive MCMC, and present examples and theorems\r\nconcerning its ergodicity and efficiency.\r\n"}