PAGE\*MERGEFORMAT29统计学习聚类方法的应用研究摘要聚类分析指将物理或抽象对象的集合分组为由类似的对象组成的多个类的分析过程.聚类分析作为一种有效的数据分析手段，能够帮助人们认识和揭示事物之间的内在联系，它已被广泛应用到数据压缩、图像处理、计算机视觉、文本聚类和异常点检测等领域.本文首先展示了统计学习的理论基础，然后详细介绍了k-均值法、基于图的谱聚类、自组织神经网络、层次聚类等聚类方法，最后使用k-均值算法来判断中国乒乓球水平以及利用谱聚类来实现图像分割等应用．关键词：k-均值法；谱聚类；自组织学习；层次聚类ResearchonApplicationofStatisticalLearningClusteringMethodABSTRACTClusteranalysisreferstotheanalysisprocessofgroupingasetofphysicalorabstractobjectsintomultipleclassescomposedofsimilarobjects.Asaneffectivemeansofdataanalysis,clusteringanalysiscanrevealtheinternalrelationshipbetweenthings.Ithasbeenwidelyusedindatacompression,imageprocessing,computervision,textclustering,outlierdetectionandotherfields.Thispaperfirstshowsthetheoreticalbasisofstatisticallearning,thenintroducesthek-meansmethod,spectrumclusteringbasedongraph,self-organizingneuralnetwork,hierarchicalclusteringandotherclusteringmethodsindetail,andfinallyusesk-meansalgorithmtojudgethelevelofChinesetabletennisandusesspectrumclusteringtoachieveimagesegmentation.Keywords：K-means;SpectralClustering;Self-OrganizedLearning;HierarchicalClustering目录TOC\o"1-2"\h\z\uHYPERLINK\l"_Toc513013835"摘要PAGEREF_Toc513013835\hIHYPERLINK\l"_Toc513013836"ABSTRACTPAGEREF_Toc513013836\hIIHYPERLINK\l"_Toc513013837"1研究意义和目的PAGEREF_Toc513013837\h11.1研究意义…......................................................................................................................................11.2研究目的..........................................................................................................................................11.3研究现状..........................................................................................................................................11.4本文主要工作..................................................................................................................................12统计学习的理论基础.......................................................................................................