|
|||||
|
|
| 基于聚类技术的集成学习方法研究 | |
| 引用本文： | 李凯,常圣领,高悦.基于聚类技术的集成学习方法研究[J].河北大学学报(自然科学版),2009,29(2). |
| 作者姓名： | 李凯 常圣领 高悦 |
| 作者单位： | 1. 河北大学,数学与计算机学院,河北,保定,071002;河北省机器学习与计算智能重点实验室,河北,保定,071002 2. 河北大学,数学与计算机学院,河北,保定,071002 3. 工业和信息化部,电信研究院,北京,100037 |
| 基金项目： | 河北省教育厅基金，河北大学博士基金 |
| 摘 要： | 研究了基于聚类技术提高分类器差异性的方法.通过Bootstrap技术与分类器学习算法训练分类器模型,利用分类器在验证集上的分类结果作为聚类的数据对象;然后应用聚类算法对这些数据聚类,并在每个簇中选择分类器代表模型,以此构成集成学习的成员;最后应用融合方法实验研究了基于聚类技术提高差异性的集成学习性能,并与集成学习方法bagging,adaboost进行了实验比较. |
| 关 键 词： | 集成学习 差异性 聚类 分类器 |
Research of Ensemble Method of Classifiers Based on Clustering Technology |
|
| LI Kai,CHANG Sheng-ling,GAO Yue.Research of Ensemble Method of Classifiers Based on Clustering Technology[J].Journal of Hebei University (Natural Science Edition),2009,29(2). | |
| Authors: | LI Kai CHANG Sheng-ling GAO Yue |
| Institution: | 1.College of Mathematics and Computer;Hebei University;Baoding 071002;China2.Key Lab of Machine Learning and Computational Intelligence of Hebei Province;China;3.China Academy of Telecommunication Research of MIIT;Beijing 100037;China |
| Abstract: | Method improving diversity of classifiers is studied based on clustering technology.In this paper,many classifiers are obtained by Bootstrap and learning algorithm of classifier.And data objects are generated by using the result for classifiers on validation set.Then clusters are obtained by using clustering algorithm on data objects and the memberships of the ensemble learning are selected from every cluster.At last,the performances of ensemble method are studied by using fused method and compared with one... |
| Keywords: | ensemble learning diversity clustering classifier |
| 本文献已被 CNKI 万方数据 等数据库收录！ | |