通用時間約束序列模式挖掘方法研究.pdf

1、沈陽化工學(xué)院碩士學(xué)位論文通用時間約束序列模式挖掘方法研究姓名：李建華申請學(xué)位級別：碩士專業(yè)：控制理論與控制工程指導(dǎo)教師：王曉峰20031231沈陽化工學(xué)院碩士學(xué)位10文Abstract摘要Sequentialpattem：miningistheprocessofdisct，’eringfrequentsequencesfromlargeamountsofdatastoredeitherindatabase，datawarehouses，

2、orotherinformationrepositoriesByimposingvariousrestrictionsonsequentialpatterns，morepracticalpatternscanbeminedSlidingwindow，maxgap，minga：、andmaxspanisreferredasuniversaltemporalconstraintssincethefourparameter：highlygen

3、eralizedifferenttemporalconstraintsInthispaperanewalgorithmrefendasMSPUispresentedtocopewithminingsequentialpatternswithuniversaltemporai：|=onstraIntsItunifiestherelatedminingalgorithmsoftemporalconstraintsequentialpatte

4、!ilsandCanminemoreprecisepatternsAllsequencesarediscoveredinonlyonedatabase、canBasedonequivalenceclassesthisalgorithmmakesuseofdivide—and—conquerstrateg?！畉odivideoriginalproblemintosmallersub—pmbtemswhichcanbeprocessedin

5、main—memoryindependentlyIneveryequivalenceclass，patternsgrowthroughthemethodsofGSlVGS2andFiIS22，whichsolvetheanti。monotonouspropertyproblemofsequencepatternswithmaxgapconstraintsIndetails，patternsgrowthprocessconsistsofs

6、tringjoin，gettingallsamesequenceidentitiesandtimeattributejoinCorrespondinglyApri、ripruning，pre—supportpmningandpost—supportpruningareintroducedintheabm。growthprocesstoremoveunfruitfulcandidatesThmestrategies，whicharebre

7、adth—firlrtsearch，depth—firstsearchandheuristicsearch，areputforwardforsearchingsequencesstaceBreaddi—firstsearchcanprunemostefficientlyanddepthfirstsearchisfittothe【listributedprocessingformininglargerdatabaseespeciallya

8、lthoughitcannotbringApriedpruningintofullplayHeuristicsearchutilizeseverypmning，whatismore，discoversmaxfrequentsequencepaftemsinadvanceandcutsdownthesequencesspacesimultaneousKeywords：datamining；sequencepatterns；aniversa