基于微博的用户兴趣-分析和个性化信息推荐

《基于微博的用户兴趣-分析和个性化信息推荐》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、Ḅᵨᡝᐶᑖ᪆ឋᓄ!"₞᤺⌕ᙠoÍḄᓝÎ=Y，ÏὶÑḄ!ÒÓÔÕÖ，@×Ø!ÙÚÛÜNᐭ!oÝÛÜ。ÞßàᩭḄI@×âS!Ḅ

2、}Ḅãä，ØåæḄ@Rçè，ᑮé¶]ê，ëᑮìᙠḄ"₞íæ。᝞eᨵᦔᙢñᵨᡝ"₞ᨵᵨḄ!，ᨬB⌕Ḅóô⁚öI᝞eᨵᦔᙢâSᵨᡝᐶ。¼÷ÑøḄsìñᡃ×úûó(Ḅᑖ᪆ᵨᡝᐶḄüᜧḄᦪ³ý，ᡂþÿÎ=ẆQḄ。?ᦻq᝞eᵨᦪ³ᑖ᪆ᵨᡝᐶ，zNOឋᓄ"₞Ḅ

3、NOᑖ᪆w¶。ìᨵḄR$，?ᦻ⌕ᨵzKÎa。✌ᐜ，ὃ⇋ᑮᩩᑁ_Ḅᱯ，ᡃ×ᨵᙠᦪ³ᵨ⚪，àIᵨ

4、᜜´᪀⚪，ᵨzqᑁ_NO，Û!"#ᙠᦪ³⚪ᦪ$a_᧕{Ḅ&⚪。ᐸ'，ᡃ×(þaIᡠᨵIᵨᡝᐶᐵḄ，!öIᡠ)Ḅ“*+”，à-.*+/qᦔV⌼ᡂ¸1。23，ᡃ×Ø4

5、☢ᑖ᪆ᵨz6*+Ḅᱯ7，᪀óὶᔠᑖ9ᘤo;ᣵ*+。ᨬÄ，ᡃ×(þᵨᡝᐶI/ÞÛ=äᓄḄ，úsÛ=ÖᩗḄ⚪ᑖ>ᩭ»?ᵨᡝᐶ。ᙠ@AY，ᡃ×Bᡃ×Ḅ4CDE▣ᑖ4wᙠᦪ³ᵨ⚪Ḅ4。@AV⊤D，?ᦻḄ4GHIᨵᦔᙢkìᵨᡝḄ@Ûᐶ。àJ，ᙠᵨᡝᦪÒKᡈὅ*+4ḄLMK，Nᯠ®zᨵᦔᙢᑖ᪆

6、sᵨᡝᐶ。——i——万方数据ᜧắᦻ᤺⌕ᐵPQ：ᐶᑖ᪆!"₞⚪LDA万方数据UserInterestAnalysisandPersonalizedInformationRecommendationBasedonMicroblogABSTRACTInthepasttenyears,theinformationontheInternetisgrowingrapidly.Wewalkfromthelackofinformationeraintotheeraofinformationexplosion.Thewaypeoplege

7、ttinginformationisalsotransforming,frommanualsearchinginformationtosearchengine,andnowtherecommendersystem.Oneofthemostimportantstepstorecommendusefulinformationtouserishowtogettheinterestofuserseffectively.Theriseofmicrobloggivesusanewandhugedatapooltoanalysisuserinterest.Itra

8、pidlybecomesaresearchhotspotinrecentyears.Inthisthesis,wetalkaboutandexploretheapproachtoanalysisuserinterestfrommicro-blogdata,andhowtomakepersonalizedrecommendation.Comparedwiththeexistingworks,thisthesishasseveraldifferences.First,consideringaboutthecharacterthatthecontentof

9、eachmicro-blogisveryshort,weconstructthetopicmodelonexternalknowledgedataset,insteadofonthemicro-blogdatadirectly,soastoenrichthesemanticsofmicro-blogcontent.Atthesametime,thisapproachavoidstheproblemofdetermineingthenumberoftopicsonmicro-blogdata.Second,weobservedthatnotallthe

10、postsarerelatedtouser’sinterest.Somepostsarethesocalled‘noisyposts’.Thesenoisypostshaveanbadimpactontheresult.Sowetalkaboutthefeaturesusedtorecognizenoisypostsfromdifferentaspects,andthenconstructacombinedclassifiertofilterthenoisyposts.—iii———万方数据ᜧắᦻABSTRACTAtlast,webe

11、lievethatuser’sinterestwillchangeovertime,sowearguetou