面向大数据的超启发式SVM的网络安全框架研究  

Research on network security framework for big data based hyper-heuristic SVM

作　　者：曹军梅[1] CAO Junmei(College of Mathematics and Computer Science,Yan’an University,Yan’an 716000,P.R.China)

机构地区：[1]延安大学计算机学院

出　　处：《重庆邮电大学学报（自然科学版）》2020年第1期23-29,共7页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金(61772406);延安大学科研引导项目(YDY2018-11)~~

摘　　要：大数据环境下,为了提高支持向量机(support vector machines,SVM)在网络安全应用环境下的性能,提出了面向大数据的超启发式SVM网络安全框架。所提超启发式SVM框架由SVM和超启发式框架组成,超启发式框架的作用是生成配置参数,并将其发送到SVM,SVM使用生成的配置来解决给定的问题,然后将成本函数发送到超启发式框架。超启发式框架分为高层策略和低层启发式,高层策略具有搜索性能,可以控制选择低层启发式并生成新的SVM配置;低层启发式算法构成了一组特定于问题的启发式算法,使用不同的规则实现对SVM配置搜索空间的探索。该框架自适应地集成了基于分解和基于Pareto方法的优点,近似SVM配置的Pareto集,解决了启发式框架的优化问题。实验结果表明,所提框架性能优于其他算法,说明框架的有效性。In order to improve the performance of Support Vector Machines(SVM)in network security applications in big data environments,network security framework for big data based hyperheuristic SVM is proposed in this paper.The proposed hyperheuristic SVM framework consists of SVM and hyperheuristic framework.The main function of the hyperheuristic framework is to generate configuration parameters and send them to SVM.SVM uses the generated configuration to solve a given problem,and then sends the cost function to the hyperheuristic framework.The hyperheuristic framework is divided into high-level strategies and low-level heuristics.High-level strategies have search performance and can control the selection of low-level heuristics and generate new SVM configuration.Low-level heuristic algorithms constitute a set of problem-specific heuristic algorithms,which use different rules to explore the search space of SVM configuration.The framework adaptively integrates the advantages of decomposition-based and Pareto-based methods,approximates the Pareto set of SVM configuration,and solves the optimization problem of heuristic framework.The experimental results show that the performance of the proposed framework is better than other algorithms,indicating the effectiveness of the framework.

关 键 词：支持向量机配置 大数据 网络安全 超启发式 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]