BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Virtual BSU Seminar: 'Assumption-lean inference for generalised li
 near model parameters' - Prof Stijn Vansteelandt\, Ghent University 
DTSTART:20200924T130000Z
DTEND:20200924T140000Z
UID:TALK150706@talks.cam.ac.uk
CONTACT:Alison Quenault
DESCRIPTION:Inference for the parameters indexing generalised linear model
 s is routinely based on the assumption that the model is correct and a pri
 ori specified. This is unsatisfactory because the chosen model is usually 
 the result of a data-adaptive model selection process\, which induces bias
  and excess uncertainty that is not usually acknowledged\; moreover\, the 
 assumptions encoded in the resulting model rarely represent some a priori 
 known\, ground truth. Standard inferences may therefore lead to bias in ef
 fect estimates\, and may moreover fail to give a pure reflection of the in
 formation that is contained in the data. Inspired by developments on assum
 ption-free inference for so-called projection parameters\, we here propose
  nonparametric definitions of main effect estimands and effect modificatio
 n estimands. These reduce to standard main effect and effect modification 
 parameters in generalised linear models when these models are correctly sp
 ecified\, but continue to capture the primary (conditional) association be
 tween two variables\, or the degree to which two variables interact (in a 
 statistical sense) in their effect on outcome\, even when these models are
  misspecified. We achieve an assumption-lean inference for these estimands
  by deriving their influence curve under the nonparametric model and invok
 ing flexible data-adaptive (e.g.\, machine learning) procedures. This talk
  aims to be broadly accessible\, focussing on concepts more than technical
 ities.
LOCATION:Virtual Seminar 
END:VEVENT
END:VCALENDAR