BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Drupal iCal API//EN
X-WR-CALNAME:Events items teaser
X-WR-TIMEZONE:America/Toronto
BEGIN:VTIMEZONE
TZID:America/Toronto
X-LIC-LOCATION:America/Toronto
BEGIN:DAYLIGHT
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
DTSTART:20190310T070000
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
DTSTART:20191103T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:664f5f6abd2c7
DTSTART;TZID=America/Toronto:20200121T100000
SEQUENCE:0
TRANSP:TRANSPARENT
DTEND;TZID=America/Toronto:20200121T100000
URL:https://uwaterloo.ca/statistics-and-actuarial-science/events/department
-seminar-lu-yang-university-amsterdam
SUMMARY:Department seminar by Lu Yang\, University of Amsterdam
CLASS:PUBLIC
DESCRIPTION:Summary \n\nDIAGNOSTICS FOR REGRESSION MODELS WITH DISCRETE OUT
COMES\n\nMaking informed decisions about model adequacy has been an outsta
nding\nissue for regression models with discrete outcomes. Standard residu
als\nsuch as Pearson and deviance residuals for such outcomes often show a
\nlarge discrepancy from the hypothesized pattern even under the true\nmod
el and are not informative especially when data are highly\ndiscrete. To f
ill this gap\, we propose a surrogate empirical residual\ndistribution fun
ction for general discrete (e.g. ordinal and count)\noutcomes that serves
as an alternative to the empirical Cox-Snell\nresidual distribution functi
on. When at least one continuous covariate\nis available\, we show asympto
tically that the proposed function\nconverges uniformly to the identity fu
nction under the correctly\nspecified model\, even with highly discrete (e
.g. binary) outcomes.\nThrough simulation studies\, we demonstrate empiric
ally that the\nproposed surrogate empirical residual distribution function
is highly\neffective for various diagnostic tasks\, since it is close to
the\nhypothesized pattern under the true model and significantly departs\n
from this pattern under model misspecification.\n
DTSTAMP:20240523T152322Z
END:VEVENT
END:VCALENDAR