Learning Pattern Classification—A Survey

Learning Pattern Classification—A Survey

ID:40374392

大小:847.39 KB

页数:29页

时间:2019-08-01

Learning Pattern Classification—A Survey_第1页
Learning Pattern Classification—A Survey_第2页
Learning Pattern Classification—A Survey_第3页
Learning Pattern Classification—A Survey_第4页
Learning Pattern Classification—A Survey_第5页
资源描述:

《Learning Pattern Classification—A Survey》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、2178IEEETRANSACTIONSONINFORMATIONTHEORY,VOL.44,NO.6,OCTOBER1998LearningPatternClassificationÐASurveySanjeevR.Kulkarni,SeniorMember,IEEE,GaborLugosi,andSantoshS.Venkatesh,´Member,IEEE(InvitedPaper)AbstractÐClassicalandrecentresultsinstatisticalpatternternclassification,stochasticcontrol,re

2、inforcementlearning,recognitionandlearningtheoryarereviewedinatwo-classclustering,etc.patternclassificationsetting.ThisbasicmodelbestillustratesWorkontheseareasspansanumberoffieldsandmanyintuitionandanalysistechniqueswhilestillcontainingtheessen-years.Inthelastseveraldecades,therewasworki

3、nthetialfeaturesandservingasaprototypeformanyapplications.Topicsdiscussedincludenearestneighbor,kernel,andhistogram1940'sand1950'sinareassuchasstatistics,informationmethods,Vapnik±Chervonenkistheory,andneuralnetworks.Thetheory,cybernetics,andearlyworkonneuralnetworksthatpresentationandt

4、helarge(thoghnonexhaustive)listofreferencesledtotremendousprogressand,infact,establishedseveralisgearedtoprovideausefuloverviewofthisfieldforbothnewfieldsofactivity.Continuedworkintheseareas,slightlyspecialistsandnonspecialists.laterworkonsystemsandcontroltheory,patternrecognition,IndexTe

5、rmsÐClassification,learning,statisticalpatternrecog-andoptimization,andmorerecentlytheexplosionofworknition,surveyreview.onneuralnetworksandothertopicssuchascomputationallearningtheory,areallpartofthisgeneralarea.I.INTRODUCTIONInthispaper,wefocusonaveryspecificsubsetofthisworkdealingwitht

6、wo-classpatternclassification.Thisproblem,HEgoaloflearningtheoryistoprovideanswerstobasicdefinedbelow,servesasaprototypeformanyreal-lifelearningTquestionssuchas:problems,whilethemathematicalsimplicityofthemodel·Whatproblemscanandcannotbelearned?allowsustogaininsightintomostofthedifficultie

7、sarising·Howmuchdataisrequired?inlearningproblems.However,thismodelbynomeans·Whataregoodalgorithmsforlearningfromexamples?coversallaspectsoflearning.Forexample,theassumptionHence,learningtheoryattemptstodelineatethefundamentalofhavingonlytwoclasseshidesmanyofthebasicdifficulties

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。