摘要
Thedatausedintheprocessofknowledgediscoveryoftenincludesnoiseandincompleteinformation.Theboundariesofdifferentclassesofthesedataareblurandunobvious.Whenthesedataareclusteredorclassified,weoftengetthecoveringsinsteadofthepartitions,anditusuallymakesourinformationsysteminsecure.Inthispaper,optimalpartitioningofincompletedataisresearched.Firstly,therelationshipofsetcoverandsetpartitionisdiscussed,andthedistancebetweensetcoverandsetpartitionisdefined.Secondly,theoptimalpartitioningofgivencoverisresearchedbythecombingandpartingmethod,acquiringtheoptimalpartitionfromthreedifferentpartitionssetfamilyisdiscussed.Finally,thecorrespondingoptimalalgorithmisgiven.Therealwirelesssignalsofftencontainalotofnoise,andtherearemanyerrorsinboundarieswhenthesedataisclusteredbasedonthetradionalmethod.Inourexperimant,theproposedmethodimprovescorrectrategreatly,andtheexperimentalresultsdemonstratethemethod’svalidity.
出版日期
2013年04月14日(中国期刊网平台首次上网日期,不代表论文的发表时间)