資源描述:
《基于加權(quán)直覺(jué)模糊集合的聚類(lèi)模型》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在教育資源-天天文庫(kù)。
1、基于加權(quán)直覺(jué)模糊集合的聚類(lèi)模型摘要:針對(duì)已有基于直覺(jué)模糊集的聚類(lèi)方法的局限性,提出了一種基于加權(quán)直覺(jué)模糊集合的聚類(lèi)模型——wifscm(clusteringmodelbasedonweightedintuitionisticfuzzysets)。在該模型中,提出了特定特征空間下的等價(jià)樣本和加權(quán)直覺(jué)模糊集合的概念;并推導(dǎo)出基于等價(jià)樣本和加權(quán)直覺(jué)模糊集合的直覺(jué)模糊聚類(lèi)算法的目標(biāo)函數(shù),利用該目標(biāo)函數(shù)推導(dǎo)出直覺(jué)模糊聚類(lèi)中心迭代算法和隸屬度矩陣迭代算法;定義了基于加權(quán)直覺(jué)模糊集合的密度函數(shù),確定了初始聚類(lèi)中心,減少了迭代次數(shù)。通過(guò)灰度圖像
2、分割實(shí)驗(yàn),證明了該模型的有效性,同時(shí)與普通直覺(jué)模糊集fcm聚類(lèi)算法(ifcm)相比,聚類(lèi)速度提高近百倍。關(guān)鍵詞:直覺(jué)模糊集;加權(quán)直覺(jué)模糊集合;聚類(lèi)中心;等價(jià)樣本;隸屬度矩陣;密度函數(shù)clusteringmodelbasedonweightedintuitionisticfuzzysetschangyan*,zhangshi.bin(schoolofnetworkengineering,chengduuniversityofinformationtechnology,chengdusichuan610225,chinaa
3、bstract:tomakeupthelimitationsofexistingclusteringmethodsbasedonintuitionisticfuzzysets,aclusteringmodelcalledwifscm(clusteringmodelbasedonweightedintuitionisticfuzzysets)isproposedbasedonweightedintuitionisticfuzzysets.inthismodel,theconceptsofequivalentsamplesandwe
4、ightedintuitionisticfuzzysetsisputforwardinspecialfeaturespace,andbasedonwhichtheobjectivefunctionofintuitionisticfuzzyclusteringalgorithmisproposed.iterativealgorithmsofclusteringcenterandmatrixofmembershipdegreeareinferredfromtheobjectivefunction.densityfunctionbas
5、edonweightedintuitionisticfuzzysetsisdefined,andinitialclusteringcenterisgottentoreduceiterativetimes.theexperimentofgrayimagesegmentationshowsthatwifscmiseffective,anditisfasterthanifcmalgorithmnearlyahundredtimes.concerningthelimitationsoftheexistingclusteringmetho
6、dsbasedonintuitionisticfuzzysets,aclusteringmodelcalledweightedintuitionisticfuzzysetmodel(wifscm)(clusteringmodelbasedonweightedintuitionisticfuzzysets)wasproposedbasedonweightedintuitionisticfuzzysets.inthismodel,theconceptsofequivalentsampleandweightedintuitionist
7、icfuzzysetwereputforwardinspecialfeaturespace,andbasedonwhichtheobjectivefunctionofintuitionisticfuzzyclusteringalgorithmwasproposed.iterativealgorithmsofclusteringcenterandmatrixofmembershipdegreewereinferredfromtheobjectivefunction.thedensityfunctionbasedonweighted
8、intuitionisticfuzzysetswasdefined,andinitialclusteringcenterwasgottentoreduceiterativetimes.theexperimentofgrayimagesegmentationsho