HansJournalofDataMining数据挖掘,2023,13(3),222-229PublishedOnlineJuly2023inHans.https://www.hanspub.org/journal/hjdmhttps://doi.org/10.12677/hjdm.2023.133022文章引用:袁梅.基于最大决策熵的快速属性约简算法[J].数据挖掘,2023,13(3):222-229.DOI:10.12677/hjdm.2023.133022基于最大决策熵的快速属性约简算法袁梅烟台大学计算机与控制工程学院,山东烟台收稿日期:2023年5月27日;录用日期:2023年6月27日;发布日期:2023年7月5日摘要在大数据时代背景下,各领域数据爆炸式增长,数据类型复杂多样。针对决策系统中基于最大决策熵的属性约简算法在大规模数据集下运行效率低的问题,提出了一种基于启发式的快速属性约简算法。本文提出的算法首先研究了属性和对象在属性约简过程中的变化对其产生影响,其次提出了属性重要度保序性的相关定理。最后通过UCI数据集对提出算法的有效性进行验证,结果表明提出的快速属性约简算法的运行效率更高。关键词快速属性约简算法,粗糙集,最大决策熵,决策系统FastAttributeReductionAlgorithmBasedonMaximumDecisionEntropyMeiYuanSchoolofComputerandControlEngineering,YantaiUniversity,YantaiShandongReceived:May27th,2023;accepted:Jun.27th,2023;published:Jul.5th,2023AbstractIntheeraofbigdata,datainvariousfieldsisgrowingexplosively,anddatatypesarecomplexanddiverse.Aimingatthelowefficiencyofattributereductionalgorithmbasedonmaximumdecisionentropyindecisionsystemunderlargedatasets,afastattributereductionalgorithmbasedonheuristicisproposed.Thealgorithmproposedinthispaperfirstlystudiestheinfluenceofthechangesofattributesandobjectsintheprocessofattributereduction,andthenputsforwardtherelatedtheoremabouttherankpreservationofattributes.Finally,theeffectivenessofthepro-posedalgorithmisverifiedbytheUCIdataset,andtheresultsshowthattheproposedfastattributereductionalgorithmismoreefficient.袁梅DOI:10.12677/hjdm.2023.133022223数据挖掘KeywordsFastAttributeReductionAlgorithm,RoughSet,MaximumDecisionEntropy,DecisionSystemCopyright©2023byauthor(s)andHansPublishersInc.ThisworkislicensedundertheCreativeCommonsAttributionInternationalLicense(CCBY4.0).http://creativecommons.org/licenses/by/4.0/1.引言粗糙集理论是用于处理不...