Rek-means
Ak-meansBasedClusteringAlgorithm
DomenicoDanieleBloisiandLucaIocchi
SapienzaUniversityofRome
DipartimentodiInformaticaeSistemistica
viaAriosto25,00185Rome,Italy
{bloisi,iocchi}@dis.uniroma1.it
doc.guandang.net
Abstract.Inthispaperwepresentanewclusteringmethodbasedonk-meansthathasbeenimplementedonavideosurveillancesystem.Rek-meansdoesnotrequiretospecifyinadvancethenumberofclusterstosearchforandismoreprecisethank-meansinclusteringdatacomingfrommultipleGaussiandistributionswithdi erentco-variances,whilemaintainingreal-timeperformance.Experimentsonrealandsyntheticdatasetsarepresentedtomeasurethee ectivenessandtheperformanceoftheproposedmethod.
Keywords:dataclustering,k-means,imagesegmentation
1Introduction
Dataclusteringtechniquesareusedtoaccomplishamultitudeoftasksinmany elds,suchasarti cialintelligence,datamining,computervision,datacompres-sion,andothers.Thereareaseriesofwellknownalgorithmsthatperformthetaskofclusteringdata:k-means,FuzzyC-means,Hierarchicalclustering,EM,andmanyothers[1].Itcanbeshownthatthereisnoabsolutebestcriterionwhichwouldbeindependentofthe nalaimoftheclustering[2].Consequently,theusermustsupplythiscriterion,insuchawaythattheresultoftheclusteringwillsuithis/herneeds.Forinstance,wecouldbeinterestedin ndingrepresen-tativesforhomogeneousgroups(datareduction),in ndingnaturalclustersanddescribetheirunknownproperties(naturaldatatypes),in ndingusefulandsuitablegroupings(usefuldataclasses)orin ndingunusualdataobjects(out-lierdetection).Whiletheabovediscussedcharacteristiciscommontoalltheclusteringmethods,itispossibletocomparetechniqueswithrespecttotimee ciencyandsolutionquality.Amongthesemethods,k-meansisthemostef- cientintermsofexecutiontime[1].However,k-meansrequirestospecifyinadvancethenumberofclustersk,itconvergesonlytoalocallyoptimalsolution,andthesolutiondependsfromtheinitializationstep.
Theaimofthispaperistopresentanextensionofk-means,calledRek-meanswhichhasthefollowingfeatures:
China Abstract—For the shortage of K-Means algorithm, a clustering algorithm based on trust value called TrustCluster is proposed, The algorithm does not ...
cient parallel clustering algorithms. In this paper, we propose a parallel k -means clustering algorithm based on MapReduce, which is a simple yet ...
International Symposium on Intelligent Information Technology Application Workshops Research on Text Clustering Algorithm Based on K_means and SOM Li Xinwu Electro...
www.elsevier.com/locate/patcog The global k-means clustering algorithm ...It is a point-based clustering method that starts with the cluster centers...
The k-means method has been shown to be effective in producing good clustering results for many practical applications. However, a direct algorithm of k-...
A PTAS for k-Means Clustering Based on Weak Coresets Dan Feldman? School ...One of the most widely used clustering algorithm is Lloyd’s 分享到: ...
2011 International Conference on Engineering and Business Management A Semi-Supervised Text Clustering Approach Based on K-Means Algorithm Lizhang Zhan1, Hong ...
(4):422—427 . Distributed K-Means clustering algorithm based on Fisher ...J Jain A K,Mur【y M N,Flynn P-J.Data clustering:a re— 翥 0 1O ...
( 2013 ) 07-0018-04 doi: 10. 3969 / j. issn. 1673-629X. 2013. 07. 005 Realization of K - means Clustering Algorithm Based on Hadoop ZHOU ...
DBSCAN VS K-means/ K-medoids ? Grid-Based & Density-based Clustering ? ...? DBSCAN is a simple and effective density-based clustering algorithm ? ...
热门文档