Social media sentiment analysis: lexicon versus machine ...
文章推薦指數: 80 %
In the present research, we compare lexicon-based and machine learning approaches to automated sentiment analysis. We aim to provide evidence of any performance. Toreadthefullversionofthiscontentpleaseselectoneoftheoptionsbelow: Otheraccessoptions YoumaybeabletoaccessthiscontentbylogginginviayourEmeraldprofile. Login RentthiscontentfromDeepDyve RentfromDeepDyve Ifyouthinkyoushouldhaveaccesstothiscontent,clicktocontactoursupportteam. Contactus Pleasenoteyoudonothaveaccesstoteachingnotes Otheraccessoptions YoumaybeabletoaccessteachingnotesbylogginginviayourEmeraldprofile. Login Ifyouthinkyoushouldhaveaccesstothiscontent,clicktocontactoursupportteam. Contactus Abstract Purpose Withthesoaringvolumesofbrand-relatedsocialmediaconversations,digitalmarketershaveextensiveopportunitiestotrackandanalyseconsumers’feelingsandopinionsaboutbrands,productsorservicesembeddedwithinconsumer-generatedcontent(CGC).These“BigData”opportunitiesrendermanualapproachestosentimentanalysisimpracticalandraisetheneedtodevelopautomatedtoolstoanalyseconsumersentimentexpressedintextformat.ThispaperaimstoevaluateandcomparetheperformanceoftwoprominentapproachestoautomatedsentimentanalysisappliedtoCGConsocialmediaandexploresthebenefitsofcombiningthem. Design/methodology/approach Asampleof850consumercommentsfrom83Facebookbrandpagesareusedtotestandcomparelexicon-basedandmachinelearningapproachestosentimentanalysis,aswellastheircombination,usingtheLIWC2015lexiconandRTextToolsmachinelearningpackage. Findings Resultsshowthetwoapproachesaresimilarinaccuracy,bothachievinghigheraccuracywhenclassifyingpositivesentimentthannegativesentiment.However,theydiffersubstantiallyintheirclassificationensembles.Thecombinedapproachdemonstratessignificantlyimprovedperformanceinclassifyingpositivesentiment. Researchlimitations/implications Furtherresearchisrequiredtoimprovetheaccuracyofnegativesentimentclassification.ThecombinedapproachneedstobeappliedtootherkindsofCGCsonsocialmediasuchastweets. Practicalimplications Thefindingsinformdecision-makingaroundwhichsentimentanalysisapproaches(oracombinationthereof)isbesttoanalyseCGConsocialmedia. Originality/value Thisstudycombinestwosentimentanalysisapproachesanddemonstratessignificantlyimprovedperformance. Keywords Sentimentanalysis Socialmedia Consumer-generatedcontent Citation Dhaoui,C.,Webster,C.M.andTan,L.P.(2017),"Socialmediasentimentanalysis:lexiconversusmachinelearning",JournalofConsumerMarketing,Vol.34No.6,pp.480-488.https://doi.org/10.1108/JCM-03-2017-2141 Publisher: EmeraldPublishingLimited Copyright©2017,EmeraldPublishingLimited × Support&Feedback Managecookies We’relistening—telluswhatyouthink Somethingdidn’twork… Reportbugshere Allfeedbackisvaluable Pleaseshareyourgeneralfeedback MemberofEmeraldEngage? Youcanjoininthediscussionbyjoiningthecommunityorlogginginhere.YoucanalsofindoutmoreaboutEmeraldEngage. Joinusonourjourney Platformupdatepage Visitemeraldpublishing.com/platformupdatetodiscoverthelatestnewsandupdates Questions&MoreInformation Answerstothemostcommonlyaskedquestionshere
延伸文章資訊
- 1Lexicon-based Detection of Violence on Social Media - Brill
Lexicon-based Detection of Violence on Social Media cognitive semantics 5 (2019) 32-69 peacefully...
- 2Dynamic polarity lexicon acquisition for advanced Social ...
Social media analytics tool aims at eliciting information and knowledge ... lexicon for augmentin...
- 3Improved lexicon-based sentiment analysis for social media ...
The algorithm consists of two key components, namely sentiment normalisation and evidence-based c...
- 4Social media sentiment analysis: lexicon versus machine ...
In the present research, we compare lexicon-based and machine learning approaches to automated se...
- 5Identifying Emotions in Social Media: Comparison of Word ...
In recent years, emotions expressed in social media messages have ... the performance of the lexi...