基于MarKov鏈的Web訪問序列挖掘算法研究.pdf

1、中南大學碩士學位論文基于MarKov鏈的Web訪問序列挖掘算法研究姓名：肖哲申請學位級別：碩士專業(yè)：軟件工程指導教師：任勝兵張輝20081118中南人學T程頌lj學位論義日錄ABSTRACTNetworkplaysamoreandmoreimportantroleinpeople’SdailylifeThereisanextensiveresearchandapplicationspacetOimproveusabilityofthew

2、ebsiteviausers’behaviorsThethesisresearchesonwebaccessingsequenceminingalgorithmbasedonMarkovchainThroughminingonewebsitelogfile，wecangetusers’behaviorpatterns，andpredicttheirpossibleaccessingpattemtothiswebsiteThiscanhe

3、lpprovidecustomizedservicestocertainusergroupOnthebasisofwebsiteusabilityanalyzingandlogfilemining，thethesisproposesanewalgorithmbasedonMarkovchainand，modifedPrefixSpansequencepatternminingalgorithmFirstly，thethesisgives

4、adeepanalysistotheclassicalsequencepatternminingalgorithmanditsresearchstatus，andalsodiscussestheproblemswhentheyareusedtominewebsequencepatternsAfterintroducingthepropertyandapplicationofMarkovchain，thethesisproposesana

5、nalyzingmethodbasedonMarkovchainAndthen，anewsequenceminingalgorithmisproposedThenewalgorithmconstructssequencedatabasewithone—stepforwardandbackwardtransitionprobabilitymatrix，andminesthedatabasethroughcombinedbi—levelan

6、dpseudoPrefixSpanprojectionAtlast，thethesisusesaninstancetoanalyzethealgorithmperformanceandverifyitseffectivenessComparedwithclassicalsequenceminingalgorithm，theproposedalgorithmcanalmostreachtheaccuracywithbigadvantage