基于熵理论及核密度估计的最大间隔学习机.pdf

第33 卷第9 期 电 子 与 信 息 学 报 Vol.33No.9 2011 年9 月 Journal of Electronics Information Technology Sept. 2011 基于熵理论和核密度估计的最大间隔学习机 刘忠宝*①② 王士同① ①(江南大学数字媒体学院 无锡 214122) ②(山西大学商务学院信息学院 太原 030031) 摘 要：该文针对支持向量机(SVM)及其变种的不足，提出一种基于熵理论和核密度估计的最大间隔学习机 MLMEK。MLMEK 引入了核密度估计和熵的概念，用核密度估计表征样本数据的分布特征，用熵表征分类的不 确定性。MLMEK 真实反映样本数据的分布特征；同时解决两类分类问题和单类分类问题；比传统SVM 具有更好 的分类性能。UCI 数据集上的实验验证了MLMEK 的有效性。 关键词：模式识别；熵理论；核密度估计；支持向量机 中图分类号：TP391.4 文献标识码： A 文章编号：1009-5896(2011)09-2187-05 DOI: 10.3724/SP.J.1146.2010.01434 A Maximum Margin Learning Machine Based on Entropy Concept and Kernel Density Estimation ①② ① Liu Zhong-bao Wang Shi-tong ① (School of Digital Media, Jiangnan Univerisity, Wuxi 214122, China) ② (School of Information, Business College of Shanxi University, Taiyuan 030031, China) Abstract: In order to circumvent the deficiencies of Support Vector Machine (SVM) and its improved algorithms, this paper presents Maximum-margin Learning Machine based on Entropy concept and Kernel density estimation (MLMEK). In MLMEK, data distributions in samples are represented by kernel density estimation and classification uncertainties are represented by entropy. MLMEK takes boundary data between classes and inner data in each class seriously, so it performs better than traditional SVM. MLMEK can work for two-class and one-class pattern classification. Experimental results obtained from UCI data sets verify that the algorithms proposed in the paper is