BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Australian Data Science Network - ECPv6.18.4.1//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-ORIGINAL-URL:https://australiandatascience.net
X-WR-CALDESC:Events for Australian Data Science Network
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:Australia/Brisbane
BEGIN:STANDARD
TZOFFSETFROM:+1000
TZOFFSETTO:+1000
TZNAME:AEST
DTSTART:20200101T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Australia/Brisbane:20210623T093000
DTEND;TZID=Australia/Brisbane:20210623T103000
DTSTAMP:20210617T012300Z
CREATED:20210617T012259Z
LAST-MODIFIED:20210617T012300Z
UID:2043-1624440600-1624444200@australiandatascience.net
SUMMARY:Bayesian hierarchical stacking—All models are wrong\, but some are somewhere useful
DESCRIPTION:Bayesian hierarchical stacking—All models are wrong\, but some are somewhere useful\nStacking is a widely used model averaging technique. Like many other ensemble methods\, stacking is more effective when model predictive performance is heterogeneous in inputs\, in which case we can further improve the stacked mixture with a hierarchical model. In this talk I will focus on the recent development of Bayesian hierarchical stacking: an approach that locally aggregates models. The weight is  a function of data\, partially-pooled\, inferred using Bayesian inference\,  and can further incorporate other structured priors and complex data. I will also discuss some theory bounds: when and why model averaging is useful; what model dissimilarity metric is relevant to Bayesian ensembles. \nRegister now
URL:https://australiandatascience.net/event/bayesian-hierarchical-stacking-all-models-are-wrong-but-some-are-somewhere-useful/
CATEGORIES:Webinar
END:VEVENT
END:VCALENDAR