微博数据挖掘研究综述.pdf

计算机研 究与发展 DOI：10．7544／issn1000—1239．2014 JournalofComputerResearchandDevelopment 51(4)：691—706，2O14 微博数据挖掘研究综述 丁兆云 贾 焰。周 斌。 ·(国防科学技术大学信息系统与管理学院 长沙 410073) z(国防科学技术大学信息系统工程重点实验室 长沙 410073) s(国防科学技术大学计算机学院 长沙 410073) (zyding@nudt．edu．cn) SurveyofDataM iningforM icroblogs DingZhaoyun，u，JiaYan。，andZhouBin。 (C0zzPgeofInformationSystemsandManagement，NationalUniversityofDefenseTechnology，Changsha410073) 。(Scienc 日dTechnologyon InformationSystemsEngineeringLaboratory，NationalUniversityofDefenseTechnology’ Changsha410073) 。(SchoolofComputer，NationalUniversityofDefenseTechnology，Changsha410073) Abstract Thepastfew yearstherapiddevelopmentandpopularizationofmicroblogshavealready beenwitnessed．Duetotheiropenness，terminalexpansion，contentsimplicity，low thresholdandSO On。microblogsdeeplyaffectourdaily lifebyprovidingan importantplatform forpeopletopublish comments，transform informationandacquireknowledge，tonamejustafew．Thoughbearingsuch advantages，microblogsmaycauseseriousimpactson thenationalsecurityandsocialdevelopmentif thevareoutofcontro1．Therefore，theresearchonmicroblogsisquitevaluablefrom boththeoretical andpracticalperspective，especiallyinthisageofthelnternet．Analyzingandminingmicroblogsalso bringsgreatchallenges．Ascanbeseen，microblogscanbetreatedasageneralizationandextensionof humanlifeinthevirtua1networkworld．However，differentfrom traditionalinformation networks， microblogshave theiruniquecharacteristics，including noisy datadiversity，socialmedia，multi— relations，therapidspreadand evolutionary，nonlinearity，largescalabilityandetc．Such differences bringforthgreatchallengesinanalyzingandminingthemicroblogs．Inthispaper，wesurveythedata miningformicroblogsandanalyzethedatasetofTwitter．Moreover，wesummarizethechallengesof dataminingformicroblogs． Keywords microblogs；datamining；textmining；socialnetwork；social