Web日志挖掘系統(tǒng)的設(shè)計(jì)與實(shí)現(xiàn).pdf

1、華中科技大學(xué)碩士學(xué)位論文Web日志挖掘系統(tǒng)的設(shè)計(jì)與實(shí)現(xiàn)姓名：唐西西申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：肖國(guó)強(qiáng)20061028華中科技大學(xué)碩士學(xué)位論文AbstractAlongwiththegradualextensionoftheWebwebsitescale，theWeblogdatainflatesincreasingly，themissionofthewebsitemanagementisgettingmoreandm

2、oredifficultHowtowithdrawalatentandusefulinformationandknowledgefromagreatdealofWeblogdata，forthepurposeofthewebsitemanagercanbeonthesegroundsimprovethestructureofthewebsite，makingaconvenienceoftheinterviewoftheuser，incr

3、easingthewebsiteinterviewquantity；HavealreadybecomeanimportantproblemforresolveurgentlycurrentlyThereforedesignandcarryoutaWeblogminingsystem，carryonexcavationtoagreatdealofWeblogdataCarriedondatacollectandpretreatmentfi

4、rst，identifiedoriginallogformatofserverandturnitsstructure，delete’’useless”recordaccordingtotheanalyticalpurposeforcarryondatadecontaminationdentifieduseraccordingtoinspiredruleandthestructureofthewebsite，thendentifyconv

5、ersationtogeteachuserconversationindifferenttimeTothenotcompleteduservisitedrecord，addedvisitedpathofuserbythestructureofthewebsite，getcompleteuserconversationPasSanalysiS，toconfirmmaximumforwardpath，todiscoverfrequentvi

6、sitedpathofeachuserByleadingKMedoidclusteringideaintothestandardgeneticalgorithm，givedanimprovedmixedgeneticclusteringalgorithm，diggedouteachdifferentuserclusterfromalluserconversationsTheimprovedalgorithmcaneffectivelyr

7、aisethesearchabilityofpartandconvergencespeedofstandardgeneticalgorithm，togetthebetterclusteringeffectMakethebestofuserclusteringresult，designakindoffastconnectionrulealgorithmofnotneedbuildawaitselectivegathertodigoutco