基于二進(jìn)制可辨矩陣的屬性約簡研究.pdf

1、分類號：干P300臣密級：——單位代碼：10359學(xué)號：——含肥工學(xué)火警HefeiUniversityofTechnology碩士學(xué)位論文MASTERDISSERTATIoNo論文題目：基于二進(jìn)制可辨矩陣的屬性約簡研究學(xué)位類別：學(xué)科專業(yè)：(工程領(lǐng)域作者姓名：同等學(xué)力計算機應(yīng)用技術(shù)汪小燕導(dǎo)師姓名：王鱟塾堡完成時間：2006年5月ResearchofattributereductionBasedonBinaryDiscernableMatr

2、ixAbstractDataMiningistheprocessofminingtheinteresting，potentiallyuseful，andunderstandableknowledgeindataClassificationisanimportantsub—branchofDataMiningandthemethodofRoughSetisoneoftheimportanttechniquesofClassificatio

3、nRoughSetisanewmathematicaltooltodealwithfuzzyanduncertainknowledgeIthasstrongknowledgeobtainingabilityItisthenewresearchdomainthatthetheoriesofRoughSetareappliedinDataMiningBecausethetheoriesofRoughSetprovidethestrictma

4、thematicsmethodoftheproblemofdealingwithdatasClassification，withouttheadditionalinformationofanydatas，theycansearchforthesmallestassemblyofthedatas，mayusedatasofdeterminingthenatureandfixingquantityandalsocangeneratether

5、uleassemblyofpolicydecisionfromthedatas，etcRoughSethasgottheextensiveapplicationNotalltheconditionattributesarenecessaryforclassification，someareunnecessaryDoingawaywiththeseattributeswon’taffectoriginalclassificationeff

6、ect，onthecontrary，itwillimprovethearticulationofthelatentknowledgeinsystemAttributereductionisreducingconditionattributesinadecisiontableThedecisiontableafterbeingreducedhasthefunctionoftheonebeforebeingreduced，butthefor

7、merhasthelessconditionattributesThedissertationismainlyresearchingonBinaryDiscernableMatrixinthetheoriesofRoughSetandrelatedtheoriesandcalculationformulasbasedontheknowledgegranulationWecancalculatethediscernmentdegree，g

8、ranulationofknowledgeandattributesignificancemakinguseofformulasobtainedWecanalsogoonattributereductionandvaluereductionofthedecisiontableutilizingrelatedformulasobtainedandpresenttweductionalgorithms：oneisattributeandit

9、svaluereductionalgorithmbasedonBinaryDiscernableMatrixOnlyscanningBinaryDiscernableMatrixonce，thealgorithmcangetcoreattributesandobjectsthatcan’tbeclassifiedcorrectly，SOwecanobtaincorevaluesReductionassembliesofcondition

10、attributescanbeacquiredifweapplytheroleofabsorptivetoeverydisjunctnormalform，SOwecangetacorevaluetablehavingattributeswhicharereducedThealgorithmmakesthatattributereductionanditsvaluereductionarecalculatedidentically，SOi