量子粒子群和最小二乘支持向量机相结合的网络异常检测

为了提高网络安全性的异常入侵检测的准确率,提出一种量子粒子群算法（QPSO）优化最小二乘支持向量机（LSSVC）的网络异常检测方法（QPSO-LSSVC）.首先利用量子粒子群处算法对LSSVC模型参数进行搜索,选出最优参数,然后采用泛化性能力优异的LSSVC对网络入侵进行建模和检测.选取KDDCUP99数据对QPSO-LSSVC性能进行测试,实验结果表明,QPSO-LSSVC提高了网络异常检测准确率,降低了误报率,为网络安全提供了有效保证.

Network Anomaly Detection by Combination of QPSO and LSSVM

In order to improve the network security intrusion detection accuracy,proposes a quantum particle swarm optimization(QPSO) optimized least square support vector machine(LSSVC) network anomaly detection method(QPSO-LSSVC).The first use of the quantum particle swarm algorithm for the parameters of LSSVC model search,optimal parameters,and then the generalization ability of excellent LSSVC modeling and detection for network intrusion.Select the KDDCUP99 data on QPSO-LSSVC performance testing,the experimental results show that,QPSO-LSSVC improves the network anomaly detection accuracy,reduce the rate of false positives,provides effective guarantee for network security.