資源描述：
《probabilistic reasoning with nave bayes and bayesian networks overview》由會員上傳分享，免費在線閱讀，更多相關(guān)內(nèi)容在教育資源-天天文庫。
1、ProbabilisticReasoningwithNa?veBayesandBayesianNetworks1ZdravkoMarkov,IngridRussellJuly,2007OverviewBayesian(alsocalledBelief)Networks(BN)areapowerfulknowledgerepresentationandreasoningmechanism.BNrepresenteventsandcausalrelationshipsbetweenthemasconditionalprobabilitiesinvolvingrandomvariables.Give
2、nthevaluesofasubsetofthesevariables(evidencevariables)BNcancomputetheprobabilitiesofanothersubsetofvariables(queryvariables).BNcanbecreatedautomatically(learnt)byusingstatisticaldata(examples).Thewell-knownMachineLearningalgorithm,Na?veBayesisactuallyaspecialcaseofaBayesianNetwork.Theprojectallowsst
3、udentstoexperimentwithandusetheNa?veBayesalgorithmandBayesianNetworkstosolvepracticalproblems.Thisincludescollectingdatafromrealdomains(e.g.webpages),convertingthesedataintoproperformatsothatconditionalprobabilitiescanbecomputed,andusingBayesianNetworksandtheNa?veBayesalgorithmforcomputingprobabilit
4、iesandsolvingclassificationtasks.ObjectivesTheaimofthisprojectistoexposestudentstotwoimportantreasoningandlearningalgorithms–Na?veBayesandBayesianNetworks,andtoexploretheirrelationshipinthecontextofsolvingpracticalclassificationproblems.Inparticular,theobjectivesoftheprojectare:?LearningthebasicsofB
5、ayesianapproachtoMachineLearningandtheBayesianNetworksapproachtoProbabilisticReasoninginAI.?Gainingexperienceinusingrecentsoftwareapplicationsintheseareasforsolvingpracticalproblems.?BetterunderstandingoffundamentalconceptsofBayesianLearningandProbabilisticReasoningandtheirrelationshipinthemoregener
6、alcontextofknowledgerepresentationandreasoningmechanismsinAI.ProjectDescription1Correspondingauthor:markovz@ccsu.edu,DepartmentofComputerScience,CentralConnecticutStateUniversity,1615StanleyStreet,NewBritain,CT06050.SimilarlytotheWebdocumentclassificationproject(http://uhaweb.hartford.edu/compsci/cc
7、li/wdc.htm)thisprojectalsohasthreemainsteps:Datacollection,Datapreparation,andMachineLearning.Thefirsttwostepsofthetwoprojectsarebasicallythesame.Infact,documentsanddatasetsinWeka’sARFFformatpreparedf
