基于微博文本的话题聚类-研究和实现

《基于微博文本的话题聚类-研究和实现》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、AbstractThedevelopmentofWeb2．0technologymakestheeraofbigdatacoming．WiththerapiddevelopmentonsocialnetworksuchaSmicro-blog,itbringsSOmanychallengesondataminingandknowledgediscovery，althoughthemicro—blogenrichesthebigdata．Comparedwiththetraditionaltextdata，themicro—blogd

2、atahaSsomedifferenceamongthepersonalinterest，entertainment，businessmarketing，andthepublicpublicity，etc．Furthermore，themicro-blogdataalsohasitsownpropertiesbothoncontentfragmentationanditsmassdata．Howtoanalyzeandmineitshidinginformationisanimportantresearchtask．Topicclu

3、steringisabasicworkonmicro—blogresearch．Byclusteringrelevantmassdataintoseveralgroupsautomatically，itsprocessingresultscanpresentsomehintsonanalyzingandminingthedata．AsthetraditionalapproachesusuallypresentSOmanyresultswithirrelevantorreplicatedinformation，itisnotfeasi

4、bletoprocesstheaboveproblemseffectively．ThetopicclusteringapproachCangrouptherelevantinformationautomatically．Furthermore，byusingthekeywordextraction，theprocessingvisualresultsareintuitional．Thisthesisdosomeresearchworksbasedonmicro—blogbyusingsomeintelligentalgorithms

5、，andthemainworksareasfollows：Firstly,wepresentsomeusefulapproachesonobtainingmicro—blogstructureddataanddatapre’processingbeforeclustering．Secondly，onthebasisofanalyzingthemicro—blogdata，wedoresearchonselectingusefulfeaturesforfurtherprocessing．Thirdly，wedesignalleffec

6、tivelyclusteringalgorithm．Onanalyzingthemicro-blogdata,wedoresearchonanalyzingwhichperformanceisbetter．Fourthly，weextractkeywordsfromclusteringresultset．ThesekeywordsCanbeusedtovisualthetopicclustering．Fifthly,wedothevisualizationprocessing，andtheresultsisclearandvisua

7、l，SOitisusefulonunderstandingandrecognitionthehiddeninformationbehindthemassdata．Theexperimentalresultsandanalysisshowsthefeasibleoftheproposedapproach．Someexistingproblemsandfurtherworksarealsopresentintheend．KeyWordsTopicclustering；Micro—blog；Featurevector；Visualizat

8、ion；InformationgainIII河北科技大学硕士学位论文IV目录摘要⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯IAbstract⋯·-⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯·⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯一I