Fast and Frugal Use of Cue Direction in States of Limited Knowledge
2000 (English)In: Proceedings of the Twenty-Second Annual Conference of the Cognitive Science Society / [ed] Lila R. Gleitman and Aravind K. Joshi, 2000, 841-846 p.Conference paper (Other academic)
An exemplar-based algorithm, PROBEX, (Juslin & Persson,1999) is shown to make robust decisions in multiple-cue inferencetasks when very few exemplars are known. We demonstratethe crucial role of knowledge of cue directions forperformance and confront PROBEX with an artificial environmentspecifically construed to favor a non-compensatoryalgorithm like Take The Best (Gigerenzer & Goldstein, 1996).PROBEX is demonstrated to perform well even in these unfavorableconditions. The explanation for the robust performanceis that PROBEX approximates Dawes Rule by using informationof the cue directions for all cues, while yet makingfew a priori assumptions about the structure of the environment.
Place, publisher, year, edition, pages
2000. 841-846 p.
Social Sciences Medical and Health Sciences
IdentifiersURN: urn:nbn:se:uu:diva-90825ISBN: 0-8058-3581-4OAI: oai:DiVA.org:uu-90825DiVA: diva2:163308
Twenty-Second Annual Conference of the Cognitive Science Society, August 13-15, 2000, Philadelphia, Pennsylvania