基于特征选择的网络入侵检测模型

研究网络安全问题,网络入侵手段多样,特征多,存在大量不利的冗余特征,传统网络入侵检测不考虑特征冗余,检测效率和正确论低。为更一步提高了网络安全,提出一种特征选择的网络入侵检测模模型。采用粒子群算法对网络系统状态特征和支持向量机参数进行同步选择,找到最优网络入侵检测模型特征和模型参数,降低了模型的输入样本维数。仿真结果表明,改进算法可降低特征维数,消除了不利于提高检测结果的冗余特征,并提高了网络入侵检测正确率,适合于小样本、实时要求高的网络入侵检测。

Network Intrusion Detection Model Based on Feature Selection

To improve network security,the paper proposed a feature selection based network intrusion detection model.Particle swarm optimization algorithm was used for the simultaneous selection of network system state features and support vector machine parameters.To find the optimal network intrusion detection model and the model parameters,the model reduced the dimensions of the input samples.The simulation results show that the algorithm can reduce the dimensions of feature space,eliminate the redundant features which are not conducive to improving the detection results,improve the detection accuracy and efficiency of network intrusion,and is suitable for application in small samples,high real-time requirement for network intrusion detection.