Decision threshold learning with incomplete feedback


Decision threshold learning with incomplete feedback

Sage Labs 4101

October 10, 2007 12:00 PM - 1:30 PM

Tom Stewart
Center for Policy Research
Nelson A. Rockefeller College of Public Affairs and Policy
University at Albany

Abstract: Many important decisions involve detection and selection. For example, who should receive extra scrutiny in airport screening and why? What conditions justify issuing severe weather warnings or terrorist-related security alerts? Which job applicants should be hired? Which college applicants should be admitted?
A common case, and the focus of this talk, is the absence of feedback about the results of negative decisions. For example, personnel managers rarely find out if applicants not hired would have been good employees. Despite the prevalence and importance of limited feedback problems in decision making, there have been only a few studies of the effects of such feedback.
A typical finding of the research that has been done, including the ongoing research to be described in this talk, is that when no feedback is given for negative decisions, people tend to make fewer positive decisions. Research results will be reviewed and models of threshold
learning will be discussed.
Paper: Decision threshold learning with incomplete feedback
Source: paper

Add to calendar