基于粗糙集的決策信息系統(tǒng)分解研究.pdf

1、分類號(hào)UDC密級(jí)編號(hào)十初大垮CENTRALSoUTHUNIVERSITY碩士學(xué)位論文論文題目一基王粗糙集姐決策信息系統(tǒng)筑解研巍學(xué)科、專業(yè)鹽篡機(jī)熬i生：復(fù)理逾研究生姓名吝崾嗍導(dǎo)師姓名及專業(yè)技術(shù)職務(wù)王加膽一敷援一ABSTRACTRoughsettheoryisallapproachofdataanalysisdealingwithimpreciseinconsistentandincompleteinformationThequantity

2、andcomplexityofdatainreal—lifeapplicationincreasecontinuouslyandroughsetdataanalysismethodbecomesinefficientAsanapproachfordatatrailsformation，decompositionofdecisioninformationsystemisalleffectivewaytoimprovetheefficien

3、cyandqualityofdataanalysisTheproblemsindecisioninformationsystemanalysisandthenecessityofdecompositiona∞discussedinthisthesis111econeeDtofdecompositionaswellasitspropertyandadvantagesaleexpatiated111eapproachesofdecision

4、informationsystemdecompositionaredividedintotwotypes：theobjectsetdecompositionapproachbasedonattributesignificanceCanimproveindependenceofclassificationandexpansibilityofrules，whiletheattributesetdecompositionapproachbas

5、edonattributedependencycanreducethecomplicatedrelationshipbetweenattributesofdifferentsub—systemsPrinciplesandalgorithmsofthetwoapproachesarepresentedandanalyzedindetailThequalityofdataanalysisandknowledgediscoverydepend

6、sondecompositionapproachcloselyComplexityofdataanalysis，accuracyofclassificationandequivalenceofdecisionareproposedascriterionsforevaluatingthedecompositionapproachesTheequivalenceofdecisioninattributesetdecompositionisa

7、nalyzedCriterionsofequivalenceontheuniverseofdecisioninformationsystemandthesamplespaceareproposedrespectivelytoavoidtheinconsistencyofclassificationandthelossofinformationcausedbydecompositionExperimentsofthedecompositi

8、onapproachesarcdesignedandimplementedUCIdatasetsaretestedandtheresultsareanalyzedaccordingtothecriterionsforevaluatingthedecompositionquality11把feasibilityandavailabilityofdecisionsystemdecompositionindataanalysisareinte