基于支持向量机和象群优化算法的入侵检测技术研究

Research on intrusion detection technology based on support vector machine and elephant herding optimization

随着信息技术的发展,网络攻击手段日渐多样化、复杂化,传统的网络入侵检测工具在网络环境中的检测准确率较低,难以应对复杂的网络攻击,因此该研究从支持向量机的分类特点出发,对象群优化算法进行改进并将其与支持向量机进行结合,建立入侵检测模型;针对入侵检测模型的参数选定性能与实际检测效果进行分析。结果显示该次设计的模型对普通攻击类型的检测正确率高达96.45%,对普通攻击类型的检测误报率低至5.72%,在面对不同攻击类型时都能保持良好的检测效果,维持良好的稳定性与鲁棒性。

With the development of information technology,the means of network attack are becoming more diversified and complex.The detection accuracy of traditional network intrusion detection tools in the existing network environment is relatively low,which is difficult to deal with complex network attacks.Therefore,in order to improve the accuracy of intrusion detection tools,this research starts from the classification characteristics of Support Vector Machine(SVM),the Elephant Herding Optimization(EHO)is improved and combined with SVM to establish the intrusion detection model.Finally,the parameter selection performance and actual detection effect of the intrusion detection model are analyzed.The results show that the detection accuracy of the designed model for ordinary attack types is as high as 96.45%,and the detection false positive rate of ordinary attack types is as low as 5.72%,showing good detection effect,good stability and robustness in the face of different attack types.