面向大规模数据集分类的增量学习方法研究.pdf

《面向大规模数据集分类的增量学习方法研究.pdf》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、分类号TP391学号14063022UDC004.9密级公开工学硕士学位论文面向大规模数据集分类的增量学习方法研究硕士生姓名谢婷婷学科专业计算机科学与技术研究方向机器学习指导教师彭宇行研究员国防科学技术大学研究生院二〇一七年二月TheResearchofIncrementalLearningmethodsforLarge-scalemulti-classDataClassificationCandidate:XieTingtingAdvisor:ProfessorYuxingPengAdissertationSubmittedinpartialfulfillmentoftherequi

2、rementsforthedegreeofMasterofEngineeringinComputerScienceandTechnologyGraduateSchoolofNationalUniversityofDefenseTechnologyChangsha,Hunan,P.R.ChinaFebruary10,2017国防科学技术大学研究生院硕士学位论文目录摘要................................................................................................................

3、........iABSTRACT...............................................................................................................iii第一章绪论............................................................................................................11.1研究背景............................................

4、.......................................................11.1.1大数据时代....................................................................................11.1.2有监督学习....................................................................................31.1.3无监督学习........................................

5、............................................41.2面临挑战...................................................................................................41.2.1有标签数据下增量学习面临的挑战............................................51.2.2无标签数据下增量学习面临的挑战............................................51.3本文工作.......

6、............................................................................................61.3.1基于随机森林的有监督增量学习方法........................................61.3.2基于自动编码器的无监督增量学习方法....................................71.4论文结构...........................................................................

7、........................8第二章相关工作....................................................................................................92.1有监督学习的增量学习方法...................................................................92.1.1Off