原创力文档- 1、本文档共17页，可阅读全部内容。
- 2、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。
- 3、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击 这里二次下载。
- 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
Na?veBayesClassifier
KeChen
2OutlineBackgroundProbabilityBasicsProbabilisticClassificationNa?veBayesExample:PlayTennisRelevantIssuesConclusions
3Background Therearethreemethodstoestablishaclassifiera)ModelaclassificationruledirectlyExamples:k-NN,decisiontrees,perceptron,SVMb)ModeltheprobabilityofclassmembershipsgiveninputdataExample:perceptronwiththecross-entropycostc)MakeaprobabilisticmodelofdatawithineachclassExamples:naiveBayes,modelbasedclassifiersa)andb)areexamplesofdiscriminativeclassificationc)isanexampleofgenerativeclassificationb)andc)arebothexamplesofprobabilisticclassification
4ProbabilityBasics Prior,conditionalandjointprobabilityforrandomvariablesPriorprobability:Conditionalprobability:Jointprobability:Relationship:Independence:BayesianRule
5ProbabilityBasics Quiz:Wehavetwosix-sideddice.Whentheyaretolled,itcouldendupwiththefollowingoccurance:(A)dice1landsonside“3”,(B)dice2landsonside“1”,and(C)Twodicesumtoeight.Answerthefollowingquestions:
6ProbabilisticClassification EstablishingaprobabilisticmodelforclassificationDiscriminativemodelDiscriminativeProbabilisticClassifier
7ProbabilisticClassification Establishingaprobabilisticmodelforclassification(cont.)GenerativemodelGenerativeProbabilisticModelforClass1GenerativeProbabilisticModelforClass2GenerativeProbabilisticModelforClassL
8ProbabilisticClassification MAPclassificationruleMAP:MaximumAPosteriorAssignxtoc*ifGenerativeclassificationwiththeMAPruleApplyBayesianruletoconvertthemintoposteriorprobabilitiesThenapplytheMAPrule
9Na?veBayes BayesclassificationDifficulty:learningthejointprobabilityNa?veBayesclassificationAssumptionthatallinputattributesareconditionallyindependent!MAPclassificationrule:for
10Na?veBayes
文档评论（0）