摘要
计算机网络攻击的多样性及隐蔽性,导致了其难以被检测,针对保护网络的安全性,准确识别网络异常问题,为了克服传统网络异常检测技术检测精度低等缺点,提出基于遗传算法优化的最小二乘支持向量机的网络异常检测方法。最小二乘支持向量机分类器(LSSVC)是一种进化的支持向量机分类器(SVC),通过构造新的二次损失函数以解决支持向量机中的二次规划问题。遗传算法用于选择合适的最小二乘支持向量机参数。选取KDDCup99数据测试采用提出的方法检测性能。实验结果表明遗传算法优化的最小二乘支持向量机分类器的网络异常检测精度高,效果好,为网络安全提供了保证。
It is difficult to detect network attack due to its diversification and concealment.In order to overcome the drawback of traditional network anomaly detection technology,such as low detection accuracy,least squares support vector machine optimized by genetic algorithm is applied to network anomaly detection in the paper.Least squares support vector machine classifier is a kind modified support vector machine classifier,which can solve the convex quadratic programming problem existing in support vector machine by constructing the new quadratic loss function,and genetic algorithm is used to adjust the parameters of least squares support vector machine classifier.KDDCUP99 experimental data are adopted to study the detection ability of the proposed method.The experimental results demonstrate that the proposed method has higher detection accuracy than normal least squares support vector machine,normal support vector machine and BP neural network.
出处
《计算机仿真》
CSCD
北大核心
2010年第12期148-151,共4页
Computer Simulation
关键词
最小二乘支持向量机
遗传算法
网络异常
检测
Least squares support vector machine（LS-SVM）
Genetic algorithm（GA）
Network anomaly
Detection
