Transcription of A Short Introduction to Boosting
{{id}} {{{paragraph}}}
Journal of Japanese Society for Artificial Intelligence, 14(5):771-780, September, 1999.(In Japanese, translation by Naoki Abe.)A ShortIntroductiontoBoostingYoav ResearchShannonLaboratory180 ParkAvenueFlorhamPark, yoav, schapire yoav, schapire a generalmethodforimprovingtheaccuracy ofany ,andexplainstheun-derlyingtheoryofboosti ng,includinganexplanationofwhyboostingof tendoesnotsufferfromover ttingaswellasboosting's relationshipto horse-racinggambler, hopingtomaximizehiswinnings,decidestocre atea computerprogramthatwillaccuratelypredict thewinnerofa horseracebasedontheusualinformation(numb erofracesrecentlywonbyeachhorse,bettingo ddsforeachhorse,etc.).To createsucha program,heasksa highlysuccessfulexpertgamblertoexplainhi sbettingstrategy. Notsurprisingly, theexpertis unabletoarticulatea grandsetofrulesforselectinga ,whenpresentedwiththedatafora speci csetofraces,theexperthasnotroublecomingu pwitha ruleofthumb forthatsetofraces(suchas, Betonthehorsethathasrecentlywonthemostra ces or Betonthehorsewiththemostfavoredodds ).
This short overview paper introduces the boosting algorithm AdaBoost, and explains the un-derlying theory of boosting, including an explanation of why boosting often does not suffer from overtting as well as boosting’s relationship to support-vector machines. Some examples of recent applications of boosting are also described. Introduction
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document:
{{id}} {{{paragraph}}}