wanghongjun
Research Associate
Supervisor of Master's Candidates
- Master Tutor
- Education Level:PhD graduate
- Degree:Doctor of engineering
- Business Address:犀浦3号教学楼31529
- Professional Title:Research Associate
- Alma Mater:四川大学
- Supervisor of Master's Candidates
- School/Department:计算机与人工智能学院
- Discipline:Electronic Information
Software Engineering
Computer Application Technology
Contact Information
- PostalAddress:
- Email:
- Paper Publications
聚类集成时机的确定
- Impact Factor:1.722
- DOI number:10.11896/j.issn.1002-137X.2015.7.011
- Affiliation of Author(s):西南交通大学
- Journal:计算机科学
- Place of Publication:重庆市
- Funded by:国家自然科学基金(61170111,61134002); 西南交通大学牵引动力国家重点实验室自主研究课题(2012TPLT15)
- Key Words:聚类集成;集成时机;泛化能力;早期停止准则;
- Abstract:使用集成学习技术可以提高聚类性能。在实验中发现,当各聚类成员聚类迭代到中后期时进行集成所得的结果会优于其迭代完全停止时进行集成所得的结果。利用集成网络泛化能力的偏差-方差分解理论对聚类集成过程中的上述现象进行解释,将提高集成网络间泛化能力的早期停止准则应用于聚类集成过程,并提出聚类集成时机的概念。对比实验表明,基于早期停止准则的聚类集成得到的结果较好,且更节约聚类集成的时间,为寻求聚类集成的最佳时机提供了可行性建议和方法。
- Co-author:wanghongjun,肖文超
- First Author:孟晓龙
- Indexed by:Academic papers
- Correspondence Author:Ellen
- Document Code:5501896
- Discipline:Engineering
- First-Level Discipline:Computer Science and Technology
- Volume:42
- Issue:07
- Page Number:48-51+84
- ISSN No.:1002-137X
- Translation or Not:no
- CN No.:50-1075/TP
- Date of Publication:2015-07-15