一种改进的PSO-SVM算法
An Improved PSO-SVM Algorithm
摘要: 粒子群优化(Particle swarm optimization, PSO)算法在整个训练数据集上搜寻支持向量机(Support vector machine, SVM)最优惩罚参数C和高斯核参数σ时会出现搜寻时间过长的问题。为了解决该问题，我们提出了一种基于距离配对排序(Distance pairing sorting, DPS)支持向量预选取的PSO-SVM算法(DPS- PSO-SVM)。该算法先将训练数据集进行DPS支持向量预选取构造一个支持向量候选集，然后利用PSO算法在支持向量候选集上对SVM参数寻优，最后将最优参数输入到SVM算法中对支持向量候选集进行训练。本文采用UCI数据库中的Breast Cancer数据和Banknote Authentication数据进行数值实验，结果表明该算法既能够缩短参数寻优时间，还能够保持PSO-SVM算法的高分类精度。
Abstract: In this paper, we introduce an improved PSO-SVM algorithm based on distance pairing sorting support vector preselecting. When particle swarm optimization (PSO) searches the optimal penalty parameter C and kernel function parameter σ of SVM on the whole training data set, the search time will be too long. In order to solve this problem, this paper proposes that the training data set uses distance pairing sorting support vector preselecting to obtain a support vector candidate set, and then the PSO parameter optimization process is put on the support vector candidate set. This can save a lot of parameter optimization time. The Breast Cancer data and Banknote Authentication data in UCI database are used in numerical experiments. The results show that the method can not only reduce the time of PSO parameter optimization, but also get good classification accuracy.
文章引用：郑恩涛, 吴思燕, 胡志涛, 鞠豪. 一种改进的PSO-SVM算法[J]. 应用数学进展, 2021, 10(7): 2305-2313. https://doi.org/10.12677/AAM.2021.107240

参考文献

[1] Vapnik, V. (1998) Statistical Learning Theory. Vol. 3, Wiley, New York, 401-492.
[2] 李航. 统计学习方法[M]. 北京: 清华大学出版社, 2012: 95-135.
[3] 杨钟瑾. 粒子群和遗传算法优化支持向量机的破产预测[J]. 计算机工程与应用, 2013, 49(18): 265-270.
[4] Refan, M.H., Dameshghi, A. and Kamarzarrin, M. (2014) Improving RTDGPS Accuracy Using Hybrid PSOSVM Prediction Model. Aerospace Science and Technology, 37, 55-69. [Google Scholar] [CrossRef] 
[5] Li, X., Yang, S.-D. and Qi, J.-X. (2006) A New Support Vector Machine Optimized by Improved Particle Swarm Optimization and Its Application. Journal of Central South University of Technology, 13, 568-572. [Google Scholar] [CrossRef] 
[6] Wang, H., Zhang, G., E, M.J. and Sun, N. (2011) A Novel Intrusion Detection Method Based on Improved SVM by Combining CPA and PSO. Wuhan University Journal of Natural Sciences, 16, Article No. 409.
[7] Wu, Q. and Law, R. (2011) Cauchy Mutation Based on Objective Variable of Gaussian Particle Swarm Optimization for Parameters Selection of SVM. Expert Systems with Application, 38, 6405-6411. [Google Scholar] [CrossRef] 
[8] 李平, 李学军, 蒋玲莉, 曹宇翔. 基于KPCA和PSOSVM的异步电机故障诊断[J]. 振动、测试与诊断, 2014, 34(4): 616-620.
[9] Al Omari, F. and Liu, G. (2015) Novel Hybrid Soft Computing Pattern Recognition System SVM-GAPSO for Classification of Eight Different Hand Motions. Optik, 126, 4757-4762. [Google Scholar] [CrossRef] 
[10] 吴开兴, 范亭亭, 李丽宏, 张琳. 基于FPCA和PSOSVM回收塑料瓶分类[J]. 计算机工程与设计, 2018, 39(11): 3555-3558+3575.
[11] Tharwat, A. and Hassanien, A.E. (2019) Quantum-Behaved Particle Swarm Optimization for Parameter Optimization of Support Vector Machine. Journal of Classification, 36, 576-598. [Google Scholar] [CrossRef] 
[12] Zhang, L. (2008) Support Vectors Pre-extracting for Support Vector Machine Based on K Nearest Neighbour Method. Proceedings of 2008 IEEE International Conference on Information and Automation, Changsha, 20-23 June 2008, 1353-1358. [Google Scholar] [CrossRef] 
[13] 韩成志, 郑恩涛, 马国春. 基于距离配对排序的支持向量预选取算法[J]. 应用数学进展, 2020, 9(2): 195-203.
[14] Schoelkopf, B. and Smola, A. (2002) Learning with Kernels. The MIT Press, Cambridge, MA.
[15] Jiao, L., Li, Y., Gong, M. and Zhang, X. (2008) Quantum-in-Spired Immune Clonal Algorithm for Global Optimization. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 38, 1234-1253. [Google Scholar] [CrossRef] 
[16] Farid, M. and Yakoub, B. (2008) Classification of Electrocardiogram Signals with Support Vector Machines and Particle Swarm Optimization. IEEE Transactions on Information Technology in Biomedicine, 12, 667-677. [Google Scholar] [CrossRef] 
[17] http://archive.ics.uci.edu/ml/datasets.php

为你推荐