Mathematical Modeling in Social Network Analysis： Using TOPSIS to Find Node Influences in a Social Network.pdfVIP

JournalofMathematicsandSystemScience3(2013)531-541 MathematicalModelinginSocialNetworkAnalysis UsingTOPSIStoFindNodeInfluencesinaSocial Network w illiam P．FoxandSeanF．Everton DepartmentofDefenseAnalysis，NavalPostgraduateSchool,Monterey,CA93943，USA Received：June13，2013／Accepted：July16，2013／Published：Octorber25，2013 Absiraet：Inasocialnetworkanalysistheoutputprovidedincludesmanymeasuresandmetrics．Foreachofthesemeasuresand metric，the outputprovidestheabilitytoobtain arankordering ofthenodesintermsofthese measures．W emightusethis informationindecisionmakingconcerningdisruptingordeceivingagivennetwork．Allisfinewhenallthemeasuresindicatethe samenodeasthekeyorinfluentialnode．Whathappenswhenthemeasuresindicatedifferentkeynodes?Ourgoalinthispaperisto exploretwomethodologiestoidentifythekeyplayersornodesinagivennetwork．WeapplyTOPSIStoanalyzetheseoutputsto findthemostinfluentialnodesasafunctionofthedecisionmkaers’inputsasaprocesstoconsiderbothsubjectivenadobjectives inputsthroughpairwisecomparisonmatrices．W eillustrateourresultsusingtwocommonnetworksfrom theliterature：theKite netw0rkna dtheInformationflownetworkfrom KnokeandWood．W ediscusssomebasicsensitivityanalysiscna maybeappliedto themethods．WefindtheuseofTOPSISasaflexiblemethodtoweightthecriterionbaseduponthedecisionmakers’inputsorthe topologyofthenetwork． Keywords：Socialnetworkanalysis，multi-attributedecisionmkaing，Analyticalhierarchyprocess(AHP)，weightedcriterion， TOPSIS，nodeinfluence Nom enclature analysis of socialnetworks in generaland dark networksinparticular[1，2]．Socialnetworkanalysisis Matrixofvaluesforalternativesbycriterion acollectionoftheoriesandmethodsthatassumesthat Matrixofnormalizedvaluesforalternativesby (rtmn criterion the b