基于微博客的突发事件情感分析的研究——以北京7.21特大暴雨为例

《基于微博客的突发事件情感分析的研究——以北京7.21特大暴雨为例》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、AbstractWiththedevelopmentofWeb2．0，theintemethasbecomeallimportanttoolofcollectinginformationforpeople．Ⅵ强thitsviralwayofspreadinginformation,charactersofshorttext，colloquialandmulti．media,microblogiSwidelyusedbypeople．Recently,moreandmoreresearchpayatt

2、entiononrealperceptivefield，commercialmarketingandsocialnetworksbasedondatafrommicroblog．TlliSpaperutilizesmicroblogAPIfromsinatocapturedata,andselectsme”Beijing721torrentialrain”astheobjectofstudyinordertoCalTyoutsentimentanalysisonmicroblogfromemerge

3、ncy．TheP印ercanbedividedintotwoparts．FirstpartiStheresearchreviewofsentimentanalysisonemergencyandthecorrespondingresearchmodel．Theotherpartistheempiricalanalysisonthe”Beijingtorrentialrain”microblogdatabasedontheoreticalresearchmodelfromformertext．Them

4、ainresearchandresultsareasfollow：1．Weanalyzethecharacteristicsandfunctionsofmicroblogdatabyliteratureresearch．Andthenwecarryoutcomparativeanalysisforthetwomainmethodsofcapturingmicroblogdata，whichincludesmicroblogAPImethodandwebcrawlermethod．2．Wemakein

5、tensivestudyonthewholeprocessofsentimentanalysis．andsummarizeitsimportantst印SwhichincludesChinesewordssegmentation，sentimentdictionaryconstruction，sentimentfeatureselectionandfeatureweightdistribution．Thenaftercomparingthedistinctivesentimentclassifica

6、tionmethodsandanalyzingthecharacteristiesoftextdata，weselectNaiveBayesModelasourclassificationmethodinempiricalanalysis3．Afterweselect”Beijingtorrentialrain"astargetevent，wetheselectedsentimentanalysismodeltoperformempificalanalysisforthemicroblogdata．

7、Underthebasisofwordssegmentationandwordsfrequencystatistics，wecarryouttheNaiveBayesProbabilityModelusingdatafromthetrainingset．Then，wemakesentimentanalysisforthetestsetusingtheaboveprobabilitymodel．4．WeusesubjectkeywordstoCalTyoutsubjectclassificationa

8、ftersentimentanalysis．Afterthat，weusecomparativemethodtoanalyzethesubjectsentimentpolarity．Thenweanalyzetheemotionalchangetrendandobjectofconcernformassmediaduringemergencyevents，i11themeanwhileweputforwardsomeconstructivesuggestionsKey