Abstract: Network attack has diversity and concealment. In order to improve the security of network abnormal intrusion detection accuracy, this paper proposes a network anomaly detection method based on Chaos Particle Swarm Optimization algorithm（CPSO） and least square support vector machine. The parameters of LSSVM are optimized by CPSO to select the optimal parameters of LSSVM, and the CPSO-LSSVM performance is tested by KDD CUP99 data. The experimental results show that the proposed method has improved the network anomaly detection accuracy, and reduced the false alarm rate. It can provide an effective guarantee for network security.

Key words: Chaos Particle Swarm Optimization algorithm（CPSO）, Least Squares Support Vector Machine（LSSVM）, network intrusion, detection

摘要： 网络攻击具有多样性和隐蔽性，为了提高网络安全性入侵检测的正确率，提出一种混沌粒子群算法（CPSO）和最小二乘支持向量机（LSSVM）相融合的网络入侵检测方法（CPSO-LSSVM）。利用混沌粒子群算法对LSSVM模型参数进行搜索，选择LSSVM最优参数，采用KDDCUP99数据集对CPSO-LSSVM性能进行测试，实验结果表明，CPSO-LSSVM提高了网络入侵检测正确率，降低了误报率，可以为网络安全提供有效保证。

关键词: 混沌粒子群优化算法, 最小二乘支持向量机, 网络异常, 检测