detect leader correl time seri queri larg graph databas graph power model tool graph data chemistri biologi imag vision social network web find maxim cliqu massiv network dgraph maxim cliqu enumer mce fundament problem graph theori import applic area social network analysi bioinformat problem extens studi exist algorithm requir memori space linear size input graph concern view massiv volum todai fastgrow network graph mce requir random access part larg graph difficult divid graph smaller part process part time result incorrect incomplet incur huge cost merg result part propos notion graph defin core network extend encompass neighborhood core mce comput propos extern memori algorithm mce extmc theh graph bound memori usag prove correct complet result comput extmc extens experi verifi extmc effici process larg network fit memori show graph captur import properti network updat maxim cliqu graph retain essenti inform low updat cost infeas perform updat entir network effici process group dorient connect queri larg graph effici discoveri frequent correl subgraph pair recent prolifer graph data wide spectrum applic led increas demand advanc data analysi techniqu view graph mine techniqu frequent subgraph mine correl subgraph mine propos applic frequenc correl plai import role paper studi problem mine set frequent correl subgraph pair simpl algorithm combin exist algorithm mine frequent subgraph correl subgraph result multipl mine oper major redund discov graph correl common graph highli correl establish theoret foundat find deriv tight lower bound correl graph correl common graph theoret result lead design effect skip mechan skip process major graph mine process algorithm fcp miner fast approxim algorithm show miss pair small set margin correl pair extens experi verifi effici effect fcp miner top dk correl graph mine effici queri process graph databas spot signific chang subgraph evolv graph correl pattern mine quantit databas effect elimin redund associ rule maintain frequent close itemset slide window abstract paper studi increment updat frequent close itemset fci slide window high speed data stream propos notion semi fci progress increas minimum support threshold itemset retain longer window drastic reduc number itemset maintain process explor properti semi fci observ major subset semi fci semi fci updat find devis effici algorithm incmin increment updat set semi fci slide window develop invert index facilit updat process empir result show incmin achiev significantli higher throughput consum memori state art stream algorithm mine fci fi incmin attain high accuraci precis recal survei algorithm mine frequent itemset data stream increas promin data stream aris wide rang advanc applic fraud detect trend learn led studi onlin mine frequent itemset fi unlik mine static databas mine data stream pose challeng addit scan natur unbound memori requir high data arriv rate data stream combinatori explos itemset exacerb mine task high complex fi mine problem hinder applic stream mine techniqu recogn critic review exist techniqu need order design develop effici mine algorithm data structur match process rate mine high arriv rate data stream unifi set notat terminolog describ paper effort main techniqu mine data stream present compr hensiv survei number state art algorithm mine frequent itemset data stream classifi stream mine techniqu categori base window model adopt order provid sight techniqu analyz algorithm exact approxim approxi mate approach fals posit fals neg discuss interest issu includ merit limit exist research substant area futur research inform dtheoret approach quantit associ rule mine abstract quantit associ rule qar mine rec ogniz influenti research problem decad due popular quantit databas us associ ation rule real life unlik boolean associ rule bar boolean attribut qar consist quantit attribut richer inform boolean attribut combin quantit attribut terval rise gener explos larg number itemset sever degrad mine effici paper propos inform theoret approach avoid reward combin attribut interv gener mine process studi mutual inform attribut quantit databas devis normal izat mutual inform make applic context qar mine strong inform relationship mine vagu associ rule abstract onlin shop applic tradit associ rule ar mine limit deal item sold ignor item sold item put basket check sold item carri hesit inform custom hesit bui hesit inform item valuabl knowledg design good sell strategi appli vagu set theori context ar mine incorpor hesit inform ar defin concept attract hesit item repres inform custom intent item base concept propos notion vagu associ rule var devis effici algorithm mine var experi show algorithm effici var captur specif richer inform tradit ar correl search graph databas correl mine gain great success applic domain abil captur underli depend object research correl mine graph databas lack fact graph data scientif domain prolifer recent year paper propos problem correl mine graph databas call correl graph search cg cg adopt pearson correl coeffici correl measur consider occurr distribut graph problem pose signific challeng subgraph graph databas candid number subgraph exponenti deriv condit set bound occurr probabl candid databas result design effici algorithm oper smaller project databas obtain significantli smaller set candid improv effici develop heurist rule appli candid set reduc search space extens experi demonstr effect method candid reduct result justifi effici algorithm mine correl larg real synthet dataset fg dindex verif dfree queri process graph databas delta dtoler close frequent itemset mine quantit correl pattern inform dtheoret approach exist research mine quantit databas focus mine associ mine associ expens practic case paper studi mine correl quantit databas show effect approach mine associ propos notion quantit correl pattern qcp found formal concept mutual inform confid devis normal mutual inform appli qcp mine captur depend attribut adopt confid qualiti measur control finer granular depend attribut specif quantit interv propos supervis method combin consecut interv quantit attribut base mutual inform interv combin guid depend attribut develop algorithm qcomin effici mine qcp util normal mutual inform confid perform level prune experi verifi effici qcomin qualiti qcp maintain frequent itemset high dspeed data stream abstract propos fals neg approach approxim set frequent itemset fi slide window exist approxim algorithm error paramet control accuraci mine result lead dilemma smaller accur mine result higher comput complex increas degrad mine accuraci address dilemma introduc progress increas minimum support function itemset retain window longer requir minimum support approach minimum support fi number potenti fi maintain greatli reduc experi show algorithm attain highli accur mine result run significantli faster consum memori exist algorithm mine fi slide window web dynam ramif develop web search engin world wide web largest hypertext system exist provid extrem rich collect inform resourc compar convent inform sourc web highli dynam factor size grow number web site page web page page content page exist hyperlink structur user search popular import tool find inform web web search engin face challeng aris web dynam paper survei research issu web dynam discuss search engin address factor web dynam briefli discuss main issu direct futur develop web search engin wuml web usag manipul languag queri web log data abstract paper develop web usag manipul languag wuml declar languag manipul web log data assum set trail form user navig process identifi web log file trail dualli model transit graph navig matrix respect underli web topolog wuml express execut transform navig log algebra nla consist sum union differ intersect project select power group oper real navig matric spars perform rang experi studi impact matrix storag scheme perform nla 