基于聚簇模型重用的概念漂移数据流半监督分类算法
Semi-supervised Classification of Data Stream with Concept Drift Based on Clustering Model Reuse
康伟 1黎利辉 1文益民1
作者信息
- 1. 广西图像图形与智能处理重点实验室(桂林电子科技大学)广西桂林 541004
- 折叠
摘要
带概念漂移的半监督数据流分类任务中,仅有少部分的数据被标记,这给分类器的训练、概念漂移的检测以及分类器对新概念的适应带来了 巨大的挑战.现有的半监督聚簇分类算法仅对分类器池中的聚簇模型进行简单的增量更新,未能有效重用历史聚簇模型.因此,文中提出了 一种新的聚簇模型重用的半监督分类算法,称为CDCMR.首先,数据流以数据块的形式到来,对数据块分完类后,训练一个簇数自适应确定的聚簇模型.其次,通过计算分类器池中的各组件分类器与聚簇模型之间的相似度,挑选多个组件分类器.再次,用当前数据块对挑选出来的组件分类器进行模型重用后,与聚簇模型集成.然后,将分类器池划分为新旧更替和多样性最大化分类器池进行更新.最后,对下一个数据块的样本进行集成分类.在多个人工和真实数据集上进行实验,结果表明,所提算法1)能有效适应概念漂移,与现有方法相比其性能有显著性提升.
Abstract
Semi-supervised classification of data stream with concept drift poses challenges to classifier training,classifier adap-tion for new concept,and concept drifting detection,for only some or even very few instances are labeled.In the existing semi-su-pervised clustering classification algorithms,only the clustering model in the classifier pool is updated incrementally,and the his-torical clustering model cannot be reused effectively.Therefore,this paper proposes a new cluster-based model reuse semi-super-vised classification algorithm,CDCMR.First,the data stream comes in the form of data chunks.After classifying the data chunks,a clustering model with adaptive determination of the number of clusters is trained.Secondly,multiple history classifiers are selected by calculating the similarity between each history classifier in the classifier pool and the clustering model.Thirdly,the selected history classifier is reused with the current data chunk and integrated with the cluster model.Then,the classifier pool is divided into old and new replacement and diversity maximization classifier pool for updating.Finally,the samples of the next data chunk are ensemble classification.Experimental results on several artificial and real data sets show that the algorithm can effec-tively adapt to concept drift,which is significantly improved compared with the existing methods.
关键词
数据流/半监督学习/概念漂移/聚簇模型重用/集成学习Key words
Data stream/Semi-supervised learning/Concept drift/Clustering model reuse/Ensemble learning引用本文复制引用
基金项目
广西壮族自治区重点研发计划(桂科AB21220023)
国家自然科学基金(62366011)
广西图像图形与智能处理重点实验室项目(GIIP2306)
出版年
2024
NETL
NSTL
万方数据