ICM优秀论文Word文档下载推荐.docx
《ICM优秀论文Word文档下载推荐.docx》由会员分享,可在线阅读,更多相关《ICM优秀论文Word文档下载推荐.docx(21页珍藏版)》请在冰点文库上搜索。
we
establish
a
model
identifyingsuspects
leader
based
on
theory
of
centricity.
We
take
three
factors
into
consideration
which
are
degree
centrality,
betweenness
centrality
and
closeness
to
evaluate
every
possible
suspect.
And
introduce
comprehensive
index
ultimately
crime
group
Gretchen.
Forproblemtwo,thispaperappliesthefuzzyevaluationmodeloftheproblemone.Itcomesoutthatthemostlikelyconspiratorsare16,81,51,56,33,57,60,28,10,79,69,13,17,15.Thecompleteprioritylistandotherdetailsofrequirement2areshowedinthepaper.
Forproblemthree,thispaperbuildsaweightedsocialnetwork---semanticnetworkonthebasisofthedemandtooptimizationthemodelwhichhasbeenestablished.Weusetherelationshipbetweenconceptsandpropertiestodefineaapplicablesemanticnetworktothecurrentcase.Thenfixingweightvaluestoeachtopicaccordingtothespecificcontentofnewsanditscontext.Finally,usingtextanalysistolearnthetopicfurthertoovercomethedefectscausedbythedataelementsinthemodel.
Atlast,thestrengthandweaknessofourmodelarediscussedandthefutureworkispointedaswell.
Keywords:
fuzzyevaluationmodel;
socialnetworkanalysis(SNA);
semanticnetwork;
textanalysis;
theory
centricity
1.Introduction
1.1Problembackground
Yourorganization,theIntergalacticCrimeModelers(ICM),isinvestigatingaconspiracytocommitacriminalact.Theinvestigatorsarehighlyconfidenttheyknowseveralmembersoftheconspiracy,buthopetoidentifytheothermembersandtheleadersbeforetheymakearrests.Theconspiratorsandthepossiblesuspectedconspiratorsallworkforthesamecompanyinalargeofficecomplex.Thecompanyhasbeengrowingfastandmakinganameforitselfindevelopingandmarketingcomputersoftwareforbanksandcreditcardcompanies.ICMhasrecentlyfoundasmallsetofmessagesfromagroupof82workersinthecompanythattheybelievewillhelpthemfindthemostlikelycandidatesfortheunidentifiedco‐conspiratorsandunknownleaders.Sincethemessagetrafficisforalltheofficeworkersinthecompany,itisverylikelythatsome(maybemany)oftheidentifiedcommunicatorsinthemessagetrafficarenotinvolvedintheconspiracy.Infact,theyarecertainthattheyknowsomepeoplewhoarenotintheconspiracy.Thegoalofthemodelingeffortwillbetoidentifypeopleintheofficecomplexwhoarethemostlikelyconspirators.AprioritylistwouldbeidealsoICMcouldinvestigate,placeundersurveillance,and/orinterrogatethemostlikelycandidates.Adiscriminatelineseparatingconspiratorsfromnon‐conspiratorswouldalsobehelpfultodistinctlycategorizethepeopleineachgroup.ItwouldalsobehelpfultotheDA’sofficeifthemodelnominatedtheconspiracyleaders.
1.2Previousresearch
Werefertosomesocialnetworkrelationmodelsaboutcrimeorganizationstructureofthearticle.Suchasthepassage‘Enterprisehiddencrimedetectionbasedonsocialnetworkanalysisoffuzzyevaluationmodel’writtenbyZengYonginChinauniversityofPetroleum,and‘criminalorganizationstructureofsocialnetworkanalysis’writtenbyChenPenginTsinghuauniversity.Wehavelearnedaboutthebasicresearchonthefuzzyevaluationmethodandrelevantevaluationparametersofthemodel.
1.3Ourwork
Weapplynetworksciencetocrimebusting.Avariablenamedconspiracylikelihoodisdefinedtoquantifyhowlikelyonepersonisaconspirator.ThenweanalyzetheCriminalinformationnetworkbyestablishingseveralmodelsfromshallowtodeep.
2.Analysis
2.1Analysisforquestionone
Thispaperestablishedafuzzyevaluationmodelbasedonsocialnetworkanalysis(SNA).Throughthesocialnetworkanalysiswecreateanindextoreflectthecharacteristicsoftheenterpriseinternalstaffsocialactivities.Usingfuzzyanalytichierarchyprocess,wecomprehensivelythinkabouttheseindexestoevaluatethepersonnelsuspectedofdegreeandthenfindoutthepossibilityofeveryonethatmayparticipateincriminalgangquantitativelyandidentifypossibleaccompliceandsortaccordingtothevolumeofsuspectedtheirpossibility.Inthelast,basedonthemodelcentricitytheoryweestablishedamodeltoidentifythesuspectleaders,gettingthemostpossiblecriminalgroupleaders.
2.2Analysisforquestiontwo
Thesolutiontoproblemtwoisthesameasthatofproblemone.Bothofthemusesocialnetworkanalysismethodtoestablishfuzzyevaluationmodel.Inthesecondquestion,weknowextraconditions.TopiconeisrelatedtotheplotandChrisisoneoftheaccomplices.
2.3Analysisforquestionthree
Semanticwebisalsoakindofnetwork,andthroughtheuseofthenetworkstructureofmutuallyconnectednodes,wetakeedgestorepresentthesemanticsoftherelationshipbetweentheconcepts.Sowecanusethetheoryofsemanticwebandacrimetheconceptsthatcanreflecttherelationshipbetweenthemultiplediagrams.Weusesemanticnetworkanalysismethodtogetherwiththeabovementionedcenterdegreeanalysisinourmodel.
3.Symbols,DefinitionsandAssumptions
3.1SymbolsandDefinitions
symbols
Definitions
i
companyemployeenumber
thecommunicationdegreeofnodesiandthesuspect
thecommunicationdegreeofnodesiandnon-suspects
u
oneofthesuspects/non-suspectentifiedsuspectsthenumberofnon-suspects
d(i,u)
theshortestdistancefromitou
theseriesofstagesrelativeLANandnodei
themostremoteplacethatnodeicanaffect
themaximumvalueoftheenterpriseLANseriesofallpersonnel
therelativenodecentralityofnodei
thenumberofreceivedmessages
thenumberofmessagessentout
thenumberofreceivedmessageswithsuspectthings
thenumberofmessagessentoutwithsuspectthings
thecenterpointofthepotential
thenumberofmessagesreceivedbythenodei
themaximumnumberofreceivedmessagesintheenterprise
3.2Generalassumptions
●Problem1,2onlythinkthatthetheme7.13.15aresuspiciousandothersarenormalsubjects,namelynotconsideringtheeffectofsubjectspecificcontentofsuspiciousdegree.Problem3requiresustotaketheconcretecontentofthethemesintoconsiderationsoastooptimizethemodelshavenbeenbuilt.
●Assumingthattheemergenceofinformationhasnothingtodowiththetime.
●Assumingthatthedataistrueanderrorscanbeignored.
●Assumingthatallmembersparticipateinthecontactandthereisnoisolatepointinthenetwork.
4.Threemodelsfortheproblems
4.1Model1:
Socialnetworkmodel
4.1.1Modelintroduction
Socialnetworkanalysisisakindofsociologicalresearchmethodsandsocialnetworkmodelisestablishedbasedonthemethod.Sociologicaltheoryisthatsocietynotmadeupofindividualsbutbythenetwork,networknodesandcontainstherelationshipbetweenthesocialnetworkanalysismethodthroughtheanalysisabouttherelationbetweenthenetworktostudythestructureandpropertyfeatureofthenetwork,includingnetworkofindividualattributesandtheoverallproperties.Networkindividualattributesanalysisincludes:
pointcentrality,closenesscentralitydegree,etc.Networkanalysisoftheoverallpropertiesincludessmallworldeffect,smallgroupresearch,condensingsubgroup,andsoon.Thismodelismainlyusedincriminology,electionpolls.
4.1.2Theestablishmentofthecriminalnetworkmodel
Intheenterprisesocialnetworkstructure,anoderepresentsacompanyemployeeandalinebetweentwonodesrepresentstheircommunication.Soweusethecollectionsofnodesanddirectedconnectionoftwonodestoestablishthemodel.
Picture1:
thestructureofcriminalnetwork
Network=(V,L)
TheNetworkmeanstherelationshipbetweendifferentnodes.Vmeansnodecollectionsorallthestaffsinthecompany.
.
isasinglenodeandNisthecollectionofnaturalnumbers.EachnodehastwokindsofcommunicationdegreeCwhicharein-degreeandout-degree.In-degreeisthenumberofmessagesthatonenodereceivedandonthecontraryout-degreeisthenumberofmessagesthatonenodesentout.Lrepresentsdifferentcommunicationpathsbetweenthenodes.
andinthisformula,
meansthenumberofmessagesthat
sentto
.TakethemethodSNAintoconsideration,basedonthenodeanditsresponserelationship,webuildaresponserelationshipmatrixS.Thereforetheenterprisecommunicationnetworkcanbeturnedintoacomputablesocialscience.
Replyrelationshipmatrix:
4.2Model2:
fuzzyevaluationmodel
4.2.1Theestablishmentofthefuzzyevaluationmodel
Thepossibilityofeachmemberintheenterprisetoparticipateinthecriminalgangisinvo