03 - Induction of Decision Trees

03 - Induction of Decision Trees

ID:39757148

大小:1.31 MB

页数:26页

时间:2019-07-10

03 - Induction of Decision Trees_第1页
03 - Induction of Decision Trees_第2页
03 - Induction of Decision Trees_第3页
03 - Induction of Decision Trees_第4页
03 - Induction of Decision Trees_第5页
资源描述:

《03 - Induction of Decision Trees》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、MachineLearning1:81-106,1986©1986KluwerAcademicPublishers,Boston-ManufacturedinTheNetherlandsInductionofDecisionTreesJ.R,QUINLAN(munnarilnswitgould.oz!quinlan@seismo,css.gov)CentreforAdvancedComputingSciences,NewSouthWalesInstituteofTechnology,Sydney2007,Australia(ReceivedAugust1,1985)Keywords:cla

2、ssification,induction,decisiontrees,informationtheory,knowledgeacquisition,expertsystemsAbstract.Thetechnologyforbuildingknowledge-basedsystemsbyinductiveinferencefromexampleshasbeendemonstratedsuccessfullyinseveralpracticalapplications.Thispapersummarizesanapproachtosynthesizingdecisiontreesthath

3、asbeenusedinavarietyofsystems,anditdescribesonesuchsystem,ID3,indetail.Resultsfromrecentstudiesshowwaysinwhichthemethodologycanbemodifiedtodealwithinformationthatisnoisyand/orincomplete.Areportedshortcomingofthebasicalgorithmisdiscussedandtwomeansofovercomingitarecompared.Thepaperconcludeswithillu

4、strationsofcurrentresearchdirections.1.IntroductionSinceartificialintelligencefirstachievedrecognitionasadisciplineinthemid1950's,machinelearninghasbeenacentralresearcharea.Tworeasonscanbegivenforthisprominence.Theabilitytolearnisahallmarkofintelligentbehavior,soanyattempttounderstandintelligencea

5、saphenomenonmustincludeanunderstandingoflearn-ing.Moreconcretely,learningprovidesapotentialmethodologyforbuildinghigh-performancesystems.Researchonlearningismadeupofdiversesubfields.Atoneextremethereareadaptivesystemsthatmonitortheirownperformanceandattempttoimproveitbyadjustinginternalparameters.

6、Thisapproach,characteristicofalargeproportionoftheearlylearningwork,producedself-improvingprogramsforplayinggames(Samuel,1967),balancingpoles(Michie,1982),solvingproblems(Quinlan,1969)andmanyotherdomains.Aquitedifferentapproachseeslearningastheacquisitionofstructuredknowledgeintheformofconcepts(Hu

7、nt,1962;Winston,1975),discriminationnets(FeigenbaumandSimon,1963),orproductionrules(Buchanan,1978).Thepracticalimportanceofmachinelearningofthislatterkindhasbeenunderlin-82J.R.QUINLANedbytheadventofknowledge-base

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。