基于复合模型的网络异常检测方法及验证

A Research on Compound Model of Network Anomaly Detection

为了提高网络异常检测的准确性，将核主成分分析与量子粒子群优化最小二乘向量机算法相结合，建立相应的网络异常检测模型。所采用的方法是：首先，通过核主成分分析对网络入侵数据进行降雏，以加快异常检测速率；然后，通过量子粒子群优化算法对最小二乘向量机进行参数优化，提高检测的准确率。仿真结果表明，复合检测模型检测提高了检测速率与准确度，为网络安全提供了保障。

In order to improve the accuracy of network anomaly detection, combined the ker- nel principal component analysis（KPCA） with quantum particle swarm optimization（QPSO） optimizated least squares vector machine （LSSVM）, a hybrid network anomaly detection model is established. KPCA is used to reduce the dimensions of network intrusion datas to speed up the detection rate, and then QPSO is applied to optimize the parameters of LSS- VM. Simulation results show that the hybrid detection model improves the rate and accuracy of detection and can provide safeguard for network security.