返回
Toward Integrating Feature Selection Algorithms for Classification and Clustering

Toward Integrating Feature Selection Algorithms for Classification and Clustering

ID：40353904

大小：830.70 KB

页数：12页

时间：2019-07-31

Toward Integrating Feature Selection Algorithms for Classification and Clustering_第1页
Toward Integrating Feature Selection Algorithms for Classification and Clustering_第2页
Toward Integrating Feature Selection Algorithms for Classification and Clustering_第3页
Toward Integrating Feature Selection Algorithms for Classification and Clustering_第4页
Toward Integrating Feature Selection Algorithms for Classification and Clustering_第5页
资源描述：

《Toward Integrating Feature Selection Algorithms for Classification and Clustering》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、IEEETRANSACTIONSONKNOWLEDGEANDDATAENGINEERING,VOL.17,NO.4,APRIL2005491TowardIntegratingFeatureSelectionAlgorithmsforClassificationandClusteringHuanLiu,SeniorMember,IEEE,andLeiYu,StudentMember,IEEEAbstract—Thispaperintroducesconceptsandalgorithmsoffeatureselection,surveysexistingfe

2、atureselectionalgorithmsforclassificationandclustering,groupsandcomparesdifferentalgorithmswithacategorizingframeworkbasedonsearchstrategies,evaluationcriteria,anddataminingtasks,revealsunattemptedcombinations,andprovidesguidelinesinselectingfeatureselectionalgorithms.Withthecateg

3、orizingframework,wecontinueoureffortstowardbuildinganintegratedsystemforintelligentfeatureselection.Aunifyingplatformisproposedasanintermediatestep.Anillustrativeexampleispresentedtoshowhowexistingfeatureselectionalgorithmscanbeintegratedintoametaalgorithmthatcantakeadvantageofind

4、ividualalgorithms.Anaddedadvantageofdoingsoistohelpauseremployasuitablealgorithmwithoutknowingdetailsofeachalgorithm.Somereal-worldapplicationsareincludedtodemonstratetheuseoffeatureselectionindatamining.Weconcludethisworkbyidentifyingtrendsandchallengesoffeatureselectionresearcha

5、nddevelopment.IndexTerms—Featureselection,classification,clustering,categorizingframework,unifyingplatform,real-worldapplications.æ1INTRODUCTIONScomputeranddatabasetechnologiesadvancerapidly,validation[18].Subsetgenerationisasearchprocedure[48],Adataaccumulatesinaspeedunmatchableb

6、yhuman’s[53]thatproducescandidatefeaturesubsetsforevaluationcapacityofdataprocessing.Datamining[1],[29],[35],[36],basedonacertainsearchstrategy.Eachcandidatesubsetisasamultidisciplinaryjointeffortfromdatabases,machineevaluatedandcomparedwiththepreviousbestonelearning,andstatistics

7、,ischampioninginturningmoun-accordingtoacertainevaluationcriterion.Ifthenewsubsettainsofdataintonuggets.Researchersandpractitionersturnsouttobebetter,itreplacesthepreviousbestsubset.realizethatinordertousedataminingtoolseffectively,Theprocessofsubsetgenerationandevaluationisrepeat

8、eddatapreprocessingisessentialtos

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。