返回
Statistical Machine Learning for Data Mining and Collaborative Multimedia Retrieval

Statistical Machine Learning for Data Mining and Collaborative Multimedia Retrieval

ID：39505460

大小：2.06 MB

页数：243页

时间：2019-07-04

Statistical Machine Learning for Data Mining and Collaborative Multimedia Retrieval_第1页
Statistical Machine Learning for Data Mining and Collaborative Multimedia Retrieval_第2页
Statistical Machine Learning for Data Mining and Collaborative Multimedia Retrieval_第3页
Statistical Machine Learning for Data Mining and Collaborative Multimedia Retrieval_第4页
Statistical Machine Learning for Data Mining and Collaborative Multimedia Retrieval_第5页
资源描述：

《Statistical Machine Learning for Data Mining and Collaborative Multimedia Retrieval》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、StatisticalMachineLearningforDataMiningandCollaborativeMultimediaRetrievalHOI,ChuHong(Steven)AThesisSubmittedinPartialFulfilmentoftheRequirementsfortheDegreeofDoctorofPhilosophyinComputerScienceandEngineeringcTheChineseUniversityofHongKongSeptember2006TheChines

2、eUniversityofHongKongholdsthecopyrightofthisthesis.Anyperson(s)intendingtouseapartorwholeofthematerialsinthethesisinaproposedpublicationmustseekcopyrightreleasefromtheDeanoftheGraduateSchool.Thesis/AssessmentCommitteeMembersProfessorTien-TsinWong(Chair)Professo

3、rMichaelR.Lyu(ThesisSupervisor)ProfessorLeoJiayaJia(CommitteeMember)ProfessorEdwardY.Chang(ExternalExaminer)Abstractofthesisentitled:StatisticalMachineLearningforDataMiningandCollaborativeMul-timediaRetrievalSubmittedbyHOI,ChuHong(Steven)forthedegreeofDoctorofP

4、hilosophyatTheChineseUniversityofHongKonginSeptember2006Statisticalmachinelearningtechniqueshavebeenwidelyappliedindataminingandmultimediainformationretrieval.Whiletradi-tionalmethods,suchassupervisedlearning,unsupervisedlearning,andactivelearning,havebeenexten

5、sivelystudiedseparately,therearefewcomprehensiveschemestoinvestigatethesetechniquesinaunifiedapproach.Thisthesisproposesaunifiedlearningparadigm(ULP)frameworkthatintegratesseveralmachinelearningtechniquesincludingsupervisedlearning,unsupervisedlearning,semi-super

6、visedlearning,activelearningandmetriclearninginasynergisticwaytomaximizetheeffectivenessofalearningtask.Basedonthisunifiedlearningframework,anovelschemeissug-gestedforlearningUnifiedKernelMachines(UKM).TheUKMschemecombinessupervisedkernelmachinelearning,unsupervis

7、edkernelde-sign,semi-supervisedkernellearning,andactivelearninginaneffec-tivefashion.AkeycomponentintheUKMschemeistolearnkernelsfrombothlabeledandunlabeleddata.Tothispurpose,anewSpectralKernelLearning(SKL)algorithmisproposed,whichisrelatedtoaquadraticprogram.Emp

8、iricalresultsshowthattheUKMtechniqueispromisingforclassificationtasks.iWithintheunifiedlearningframework,thisthesisfurtherexplorestwoimportantchallengingtasks.OneisBatchModeAc

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。