计算机应用研究2017，Vol.34Issue(11)：3233-3236,4.DOI:10.3969/j.issn.1001-3695.2017.11.007
基于动态权重的AdaBoost算法研究
Research on dynamic weights based AdaBoost
摘要
Abstract
Ensemble learning algorithms such as AdaBoost combine base classifiers statically so that they cannot customize proper combinations of base classifiers for different testing samples.To deal with this drawback,this paper proposed a new ensemble learning algorithm which weighted base classifiers dynamically.In the training phase,the algorithm grouped training samples into clusters and associated each base classifiers with a weight that decided by the fitness between base associated and clusters.In the testing phase,it calculated the weights of base classifier on a testing sample using the distance between the testing sample and the clusters.Experiment results on UCI datasets show that the proposed algorithm obtains better performance than traditional ensemble learning algorithms by exploring the diversity among testing samples.关键词
AdaBoost/动态权重/聚类/基分类器Key words
AdaBoost/dynamic weights/clustering/base classifier分类
信息技术与安全科学引用本文复制引用
张亮,李智星,王进..基于动态权重的AdaBoost算法研究[J].计算机应用研究,2017,34(11):3233-3236,4.基金项目
国家自然科学基金青年基金资助项目(61502066) （61502066）
重庆市基础与前沿研究计划项目(cstc2015jcyA40018) （cstc2015jcyA40018）
重庆市教委科学技术研究项目(KJ500438) （KJ500438）
重庆市研究生科研创新资助项目(CYS15167) （CYS15167）
