基于改进SGO算法优化的SVM网络入侵检测  

SVM Network Intrusion Detection Optimized by an Improved SGO Algorithm

作　　者：张小萍[1] 李秋兰 ZHANG Xiaoping;LI Qiulan(College of Computer and Electronic Information,Guangxi University,Nanning 530004,China)

机构地区：[1]广西大学计算机与电子信息学院,广西南宁530004

出　　处：《云南师范大学学报(自然科学版)》2025年第1期49-55,共7页Journal of Yunnan Normal University:Natural Sciences Edition

基　　金：广西数字基础设施重点实验室资助项目(GXDINBC202401).

摘　　要：利用改进社会群体优化(SGO)算法优化支持向量机(SVM)算法的惩罚系数c和核函数g,从而提高网络入侵检测的准确率;同时为了缓解SGO算法存在的随机初始化不均匀、陷入局部最优等问题,在社会群体优化算法的初始化阶段加入佳点集使得初始化种群更加均匀;在SGO算法的获得阶段加入黄金正弦算法使其跳出局部最优,进而有效地提升SVM分类模型的准确率.利用KDD99数据集进行仿真实验,实验证明提出算法具有检测时间短、准确率较高和误报率低的优势.An improved social group optimization(SGO)algorithm was employed to optimize the penalty coefficient c and kernel function g of the support vector machine(SVM)algorithm,thereby enhancing the accuracy of network intrusion detection.To address issues inherent in the SGO algorithm,such as uneven random initialization and convergence to local optima,good point set was incorporated during the initialization phase of the social group optimization process.This adjustment ensured a more uniform initial population.Additionally,during the acquisition phase of the SGO algorithm,a golden sine algorithm was introduced to facilitate escaping from local optima,which effectively improved the accuracy of the SVM classification model.Simulation experiments were conducted using the KDD99 dataset.The results demonstrated that the proposed algorithm offered advantages in terms of shorter detection times,higher accuracy rates,and lower false positive rates.

关 键 词：社会群体优化算法 支持向量机 入侵检测 KDD99 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]