期刊文献+

数据集动态重构的集成迁移学习 被引量:5

Ensemble transfer learning algorithm based on dynamic dataset regroup
下载PDF
导出
摘要 目前很多数据挖掘和机器学习方法都有一个基本假设:训练数据和测试数据必须服从相同的分布。但是在很多情况下这种假设不成立,没有考虑分布差异的传统机器学习方法就不能正确分类了。提出了一种新的迁移学习方法DRTAT,对原训练数据进行动态分割重组,适时地淘汰冗余数据,并进行分类器的集成。通过在多个文本数据集和UCI数据集上进行测试,并与TrAdaboost算法进行比较,表明了算法的先进性。 There is a basic assumption in many existing data mining and machine learning techniques,that training and test data must be governed by the same distribution.However,this assumption does not hold in many cases,then traditional machine learning methods not aware of the difference of distribution may fail.This paper proposes a novel transfer-learning algorithm called DRTAT,which dynamically regroups the primary training data sets and eliminates the redundancy data timely,then makes classifiers ensemble.The experiments are performed on many text data sets and the UCI benchmark data sets,and DRTAT is compared with TrAdaboost algorithm,the results show the superiority of DRTAT.
作者 刘伟 张化祥
出处 《计算机工程与应用》 CSCD 北大核心 2010年第12期126-128,共3页 Computer Engineering and Applications
基金 山东省中青年科学家科研奖励基金(博士基金)(No.2006BS01020) 山东省高新技术自主创新工程专项计划(No.2007ZZ17) 山东省自然科学基金No.Y2007G16 山东省科技攻关计划No.2008GG10001015 山东省教育厅科技计划项目No.J07YJ04~~
关键词 分布差异 知识迁移 动态数据集重组 冗余数据淘汰 分类器集成 distribution difference knowledge transfer dynamic dataset regroup eliminating the redundancy data classifier ensemble
  • 相关文献

参考文献8

  • 1Dai Wen-yuan,Yang Qiang,Xue Gui-rong,et al.Transferring Naive Bayes classifiers for text classification[J].Association for the Advance-ment of Artificial Intelligence,2007:540-545.
  • 2Daum'eIII H,Marcu D.Domain adaptation for statistical classifiers[J].Journal of Artificial Intelligence Research,2006,26:101-126.
  • 3Torrey L,Shavlik J,Natarajan S,et al.Transfer in reinforcement learn ing via Markov logic-networks[J].Association for the Advancement of Artificial Intelligence,2008.
  • 4Dai Wen-yuan,Yang Qiang,Xue Cui-rong,et al.Boosting for trana-fer learning[C]//ACM International Conference Proceeding Series,2007,227:193-200.
  • 5Li Cen.Classifying imbalanced data using a bagging ensemble varia-tion(BEV)[C]//ACM Southeast Regional Conference,2007:203-208.
  • 6Witten I H,Frank E.数据挖掘实用机器学习技术[M].2版.北京:机械工业出版社,2006:212-215.
  • 7Freund Y,Schapire R E.A decision theoretic generalization of on-line learning and an application to boosting[J].Journal of Computer and System Sciences,1997,55(1):119-139.
  • 8Mitchell T M.机器学习[M].北京:机械工业出版社,2006:166-167.

共引文献3

同被引文献71

  • 1李秋洁,茅耀斌,叶曙光,王执铨.代价敏感Boosting算法研究[J].南京理工大学学报,2013,37(1):19-24. 被引量:3
  • 2李鹏飞,吴太成.桥梁健康监测技术研究综述[J].建筑监督检测与造价,2010(7):24-27. 被引量:3
  • 3L Rigutini, M Maggini, B Liu. An EM based training algorithm for cross-language text categorization [ A ]. IEEE International Conference on Web Intelligence[ C ]. University of Technology of Compiegne, France,2005.282- 287.
  • 4W Dai,Q Yang, G-R Xue,Y Yu. Boosting for transfer learning A ]. Proceedings of the Twenty-Fourth International Conference on Machine Learning[C]. Orvallis, Oregon, USA, 2007. 193 - 200.
  • 5W Dai, Y Chen, G-R Xue, Q Yang, Y Yu. Translated learning: Transfer learning across different feature spaces[A]. Advances in Neural Information Processing Systems 21 [C]. Vancouver, British Columbia, Canada, 2009.786 - 791.
  • 6Y Liu,P Stone. Value-function-based transfer for reinforcement learning using structure map-ping [ A ]. Proceedings of the Twenty-First National Conference on Artificial Intelligence [ C]. Boston, Massachusetts, 2006.877 - 882.
  • 7Sinno J Pan, Qiang Yang. A survey on transfer learning[ A ]. IEEE Transactions on Knowledge and Data Engineering [ C ]. Los Alamitos, CA, USA, 2009.556 - 562.
  • 8R Raina,A Battle,H l_ee,B Packer,A Y Ng. Self-taught learning: Transfer learning from unlabeled data[A]. The Twenty-fourth International Conference on Machine Learning[ C]. Corvallis, Oregon, USA, 2007.759 - 766.
  • 9W Dai, G.-R. Xue, Q Yang, Y Yu. Transferring naive bayes classifiers for text classification[A]. The Twenty-Second National Conference on Artificial Intelligence[ C ]. Corvallis, Oregon, USA, 2007. 540 - 545.
  • 10戴文渊.基于实例和特征的迁移学习算法研究[D].上海:上海交通大学,2010.6.

引证文献5

二级引证文献35

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部