基于類(lèi)別相似性和分類(lèi)方法的個(gè)性化推薦方法研究.pdf

1、燕山大學(xué)碩士學(xué)位論文基于類(lèi)別相似性和分類(lèi)方法的個(gè)性化推薦方法研究姓名：湯顯申請(qǐng)學(xué)位級(jí)別：碩士專(zhuān)業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：郭景峰20051201燕山大學(xué)工學(xué)碩士學(xué)位論文AbstractWiththedevelopmentofintemetandE。commercetherecommendation、systemhasgraduallybecomeartimportantresearchfieldofEcommercetechnology

2、andattractsmanyresearchers’attentionCollaborativefilteringtechnologyandcategory—basedrecommendationtechnologyarethemainresearchissuesinthefieldofpersonalizedrecommendation，andinthispaperwefoCUSourresearchonthesetwoissues

3、Inthefieldofcollaborativefilteringrecommendationtechnologytheproblemsthatexistintoday’ScoIlaborativefilteringtechnologyhavebeenanalyzedandpointedoutthatwiththedevelopmentofE—commerce，themagnitudesandcommoditiesgrowrapidl

4、ywhichresultedintheextremesparsityofuserrationdataFornotbeingtakenthecategoryofdifferentitemsintoaccount，thetraditionalsimilaritymeasuremethodsworkinaccuratelyinthissituationAimingatthisproblem，anewmethodhasbeenbroughtfo

5、rwardtodeal麗thusersratingdatainordertoimprovetheveracityofitems’similarityWhencomputingtheinitialforecastratingofitems，anewmethodisintroducedtoimprovethedatadensityofusersratingmatrixAndatthefinalforecastratingoftheitems

6、，anewmethodispresentedtoimprovetheveracityofusersforecastratingInordertoimprovethecharacterofrealtime，anincrementalcollaborativefilteringrecommendationalgorithmbasedoncategorysimilarityispresentedTheexperimentalresultssh

7、owthatthen鉀VmethodsCanefficientlyresolvetheexistedproblemoftraditionalmethodsintheconditionoftheextremesparsityofuserratingdata，andthequalityoftherecommendedresultcan_alsobeefficientlyimprovedInthefieldofcategory—basedre