Treść książki
Przejdź do opcji czytnikaPrzejdź do nawigacjiPrzejdź do informacjiPrzejdź do stopki
Abstract
TheWeb2.0revolutionspreadingovertheInternethasdramaticallychanged
thewaydataisgatheredandprocessedbywebapplications.Thestatic,au-
thoritarianmodeloftheWebhasbeenabandonedinfavorofdynamic,
community-drivenmodelofuser-generatedcontent.Socialnetworksappear
abundantlyinalldomainsofhumanactivity,presentinguserswithlimitless
volumesofdata,information,andknowledge.Unfortunately,unearthingthe
knowledgehiddeninvastrepositoriesofsocialapplications,suchaswikis,
Internetforums,ortheblogosphere,isadifficultandchallengingtask.Struc-
turalcomplexity,hugevolumeofdatatobeprocessed,stochasticnatureof
socialprocessesunderlyingthedata,allcontributetothehardnessofthis
task.Dataminingmethodsdevelopedforrelationaldatarepositoriescan-
notbesimplyadaptedtosocial-drivendata.Newmodelsandalgorithms
arerequiredfordiscoveringknowledgeinsocial-drivendata.
Thisdissertationintroducesatrust-basedapproachforminingsocial-
drivendata.Theauthorexaminesdifferenttypesofsocial-drivendata,in-
cludingblogs,Internetforums,andonlineauctions,andutilizescommon
underlyingnotionsoftrustandcredibilitytodevelopalgorithmsformining
social-drivendata.Usingthenotionsoftrustandcredibilityallowstodis-
covervariousimportantpatternsinsocial-drivendata.Intheblogosphere,
trustmanifestsitselfastherankingofblogsbasedontheirrelativeinflu-
enceontheblogosphere.InthedomainofInternetforums,miningthesocial
networkofparticipantsunveilstruesocialrolesattributedtoparticularpar-
ticipants.Finally,trustandcredibilityformthefoundationofreputation
modelsfortheparticipantsofonlineauctions.
Thedissertationpresentsnewmodelsandalgorithmsforminingsocial-
drivendata.Allalgorithmshavebeenimplementedandtheireffectiveness
hasbeenverifiedbythoroughexperiments.Theresultsoftheexperimental
evaluationofmodelsandalgorithmsallowedtoconfirmthemainthesisof
thedissertation,namely,thattrustandcredibilitywerethemostimpor-
tantandcrucialnotionsusedtocreaterelationshipsinsocial-drivendata.
Inaddition,ithasbeenproventhattrustandcredibilitymightbediscov-
eredautomaticallybyusingdataminingmethodsontheunderlyingsocial
networks.
