作者：王建国, 张鑫礼, 张文兴

作者单位：内蒙古科技大学机械工程学院, 内蒙古 包头 014010

关键词：支持向量机;模糊C均值聚类;粒度计算;粒度支持向量机;核方法

摘要：

针对传统粒度支持向量机(granular support vector machine,GSVM)在处理大规模数据集时划分方法的随机性严重影响模型训练效能的情况,提出一种基于核模糊C均值聚类的粒度支持向量机(granular support vector machine based on kernel-based fuzzy c-means cluster, GSVM-KFCM)的方法。首先利用核映射将数据映射到高维空间进行聚类划分得到若干个信息粒,然后在每个信息粒中进行支持向量机的训练,提取出关键信息并融合建立最终决策模型。实验结果表明:该方法可以降低大规模数据集的训练时间,同时也能提高算法的准确度。

Granular support vector machine based on kernel-based fuzzy C-means cluster

WANG Jianguo, ZHANG Xinli, ZHANG Wenxing

School of Mechanical Engineering, Inner Mongolia University of Science and Technology, Baotou 014010, China

Abstract: The training efficiency of models is often seriously affected by the granulating randomness of traditional granular support vector machines(GSVM) when computing with large-scale data sets. A new GSVM based on Kernel-based Fuzzy C-Means Cluster(GSVM-KFCM) has been proposed to solve this problem. First, GSVM-KFCM was used to map the original data into a high dimension space and then split them into several information granules, each of which was trained with the support vector machine(SVM), and crucial information were extracted and combined to build up a final decision-making model. The experimental results have proved that this new method can reduce the training time of large-scale data sets and can also improve the accuracy to some extent.

Keywords: support vector machine;fuzzy C-means cluster;granular computation;granular support vector machine;kernel-based method

2016, 42(2): 96-99  收稿日期: 2015-2-17;收到修改稿日期: 2015-4-25

基金项目: 国家自然科学基金(21366017) 内蒙古自然科学基金重大项目(2011ZD08)

作者简介: 王建国(1958-),男,内蒙古呼和浩特市人,教授,硕士生导师,博士,研究方向为机电系统智能诊断与复杂工业过程建模、优化及故障诊断制。

参考文献

[1] VAPNIK V. The nature of statistical learning theory[M].New York:Springer-Verlay Press,1995:156.
[2] TANG Y H, JIN B, SUN Y, et al. Granular support vector machines for medical binary classification problems[M].Computational Intelligence in Bioinformatics and Computational Biology,2004(4):73-78.
[3] 张丽娟,李舟军,陈火旺. 粒度计算及其在数据挖掘中的应用[J]. 计算机科学,2005,32(12):178-180.
[4] DING S F, QI B J. Research of granular support vector machine[J]. Artif Intell Rev,2012,38(5):1-7.
[5] TANG Y C, JIN B, ZHANG Y Q. Granular support vector machines with association rules mining for protein homology prediction[J]. Artificial Intelligence in Medicine, 2005(35):121-134.
[6] 王文剑,郭虎升. 粒度支持向量机学习模型[J]. 山西大学学报(自然科学版),2009,32(4):535-540.
[7] KE L, HUANG J. Study on a GA-based SVMdecision tree multi-classiflcation strategy[J]. Acta Electron Sinica,2008,36(8):1502-1507.
[8] 段丹青,陈松乔,杨伟军,等. 使用粗糙集和支持向量机检测入侵[J]. 小型微型计算机系统,2008,29(4):627-630.
[9] 郭虎升,王文剑. 基于神经网络的SVM学习算法[J]. 计算机工程与应用,2009,45(2):51-54.
[10] 文贵华,向君,丁月华. 基于商空间粒度理论的大规模分类算法[J]. 计算机应用研究,2008,25(8):2299-2301.
[11] MARK G. Mercer kernel-based clustering in feature space[J]. IEEE Transactions on Neural Networks,2002, 13(3):780-784.