适于渐变概念漂移数据的自适应分类算法 - 模式识别与人工智能.pdf

第26 卷摇 第7 期摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 模式识别与人工智能摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 Vol.26摇 No.7 摇 2013年7月摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 PR AI摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 Jul摇 2013 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 适于渐变概念漂移数据的自适应分类算法* 1,2 1 1 张景祥 摇 摇 王士同摇 摇 邓赵红 1(江南大学 数字媒体学院摇 无锡214122) 2(江南大学 理学院摇 无锡214122) 摘摇 要摇 数据的概念漂移特性是广泛存在的.针对渐变概念漂移的分类问题,提出一种自适应近邻投影均值差支 持向量机算法.该算法基于结构风险最小化模型,以再生核Hilbert空间中近邻投影均值差为相邻分类器间差异的 度量,在全局优化中融入数据自身的分布特征,提高算法的适应性.在模拟数据和真实数据集上的实验结果表明该 算法是有效的. 关键词摇 渐变概念漂移, 自适应分类,支持向量机 中图法分类号摇 TP391.4 Adaptive Classification Algorithm for Gradual Concept鄄Drifting Data 1,2 1 1 Zhang Jing鄄Xiang ,Wang Shi鄄Tong ,Deng Zhao鄄Hong 1(School of DigitalMedia,Jiangnan University,Wuxi214122) 2(School of Science,Jiangnan University,Wuxi214122) ABSTRACT At present, the concept鄄drifting phenomena in various datasets receives considerable attention. Aiming at the classification of concept drift,an adaptive neighbor projection mean discrepancy support vector machine (NMD鄄SVM) is proposed. The concept of the neighbor projection mean discrepancy in the reproducing kernel Hilbert space is defined to measure the discrepancy between adjacent sub鄄classifiers,and the distribution characteristics of data are integrated into the process of global optimization. Thus,the adaptability of classification algorithm is enhanced. The experimental results on the simulation and real datasets validate the effectiveness of the proposed algorithm. Key Words摇 Gradual Concept Drift,Adaptive