波段性网络入侵差异化特征的有效提取模型仿真

发布时间：2018-01-05 01:12

本文关键词：波段性网络入侵差异化特征的有效提取模型仿真　出处：《科技通报》2015年02期 　论文类型：期刊论文

【摘要】：在对波段性网络入侵差异化特征进行提取的过程中,会出现入侵差异化特征伪装程度逐渐升高的情况,导致传统的基于敏感性数据挖掘的波段性网络入侵差异化特征提取方法,由于不能有效区分入侵特征与正常特征,无法有效实现波段性网络入侵差异化特征的有效提取,提出一种基于支持向量机的波段性网络入侵差异化特征提取模型,获取不确定入侵中波段性网络节点的差异化特征以及入侵节点,给出两种不确定入侵中波段性网络节点的差异化特征训练数据样本集,通过非线性映射将数据样本集从原空间映射到高维特征空间中,得到高维特征空间中最优线性分类面,采用支持向量机求解该分类面的优化解,使用网格搜索法,通过调整错分惩罚因子与核宽度,分别训练不同的支持向量机,获取泛化能力最强的参数组合,完成入侵中波段性网络节点的差异化特征所对应的数据集的核参数优化和分类,实现波段性网络入侵节点的差异化特征的有效提取。仿真实验结果表明,所提方法具有很高的准确性及有效性。
[Abstract]:The extraction process in the band of the different characteristics of network intrusion, intrusion will appear different characteristics of camouflage degree gradually increased, leading to the sensitivity of data mining the band network intrusion feature extraction method based on the traditional difference, because it cannot effectively distinguish between the invasion characteristics and normal characteristics, can not effectively achieve the effective extraction of network band the invasion of different characteristics, this paper proposes a support vector machine band of network intrusion feature extraction model based on difference, get the differentiation features of the invasion of the band of network nodes and node intrusion, gives two kinds of uncertain difference in band network intrusion feature training data set, through nonlinear mapping data set from the original space to a high dimensional feature space, and obtain the optimal linear high-dimensional feature space classification, using support Vector machine to solve the classification of the optimization solution, using the grid search method, penalty factor and kernel width by adjusting the wrong points respectively, support vector machine training different parameters, obtaining best generalization ability, complete nuclear parameter optimization and classification of the corresponding differentiation features of the invasion of the band of network nodes in the data set, effective extraction of differentiated characteristics of the band of network intrusion node. Simulation results show that the proposed method has high accuracy and validity.

【作者单位】：琼州学院电子信息工程学院;
【分类号】：TP393.08
【正文快照】： 0引言随着波段性网络技术的不断发展以及恶意入侵的不断增多,波段性网络的安全问题已逐渐影响了人们日常生产、生活的诸多方面[1,2]。所以,波段性网络入侵差异化特征检测方法,已经成为波段性网络领域需要研究的重点问题,受到了很多专家的广泛关注[3-5]。在对波段性网络入侵差

【参考文献】