變精度模糊粗糙集理論與應用研究.pdf

1、分拳!號：UDC：Y1436094密級：——編號：——理學碩士學位論文變精度模糊粗糙集理論與應用研究碩士研究生：指導教師：學位級別：學科、專業(yè)：所在單位：論文提交日期：論文答辯日期：學位授予單位：張月童鄧廷權教授理學碩士應用數(shù)學理學院2008年5月2008年6月哈爾濱工程大學哈爾濱T程大學碩十學何論文ABSTRACTAsanewtheoryofdataanalysistheroughsetsareallimportantmathemat

2、icalmethodtodealwithuncertaintyproblemsItiswidelyappliedinvariousfieldssuchasdatamining，machinelearning，artificialintelligenceandpatternrecognitionButtherearesomeshortcomingsaboutclassicalroughsetstheoryPawlak’Sroughsets

3、modelandthevariableprecisionroughsetsmodelarebuiltonequivalencerelationsButinpracticalapplications，manyobjectswehandlearefuzzy，andtherearenocrisprelationsbetweenoneobjectandanotherbutonlyfuzzyrelationsAttributesreduction

4、isaNPproblem，andtheconventionaloptimizationCan’tbeusedsuccessfullyInthisthesistheACOalgorithmisusedtosolvethisproblemOnthebasisofthefeatureofattributesreduction，thestatetransitionprobabilityformulaandthepheromoneupdatafo

5、rmulaaremodifiedtomakeitmoreadaptivetoattributesreductionThisthesisproposesthevariableprecisionfuzzyroughsetsmodelonthebasisoffuzzysimilarityrelations，whichassimilatestheideaofZiarko’SvariableprecisionroughsetsThentherou

6、ghapproximationsoffuzzysetsarestudied，andparticularlythepropertiesofvariableprecisionfuzzyroughsetsareinvestigatedPawlak’SroughsetsandtheZiarko’SvariableprecisionroughsetsaretwoinstancesofthemodelproposedInclusteringanal

7、ysis，differentattributeshavedifferentimportanceinclustering，andevensomeattributesdisturbtheclusteringresultsInthisthesis，thevariableprecisionfuzzyroughsetsmodelisappliedincalculatingtheweightsofattributesTheweightsofattr