基于特征选择的网络入侵检测模型研究

为了有效从收集的恶意数据中选择特征去分析,保障网络系统的安全与稳定,需要进行网络入侵检测模型研究。但目前方法是采用遗传算法找出网络入侵的特征子集,再利用粒子群算法进行进一步选择,找出最优的特征子集,最后利用极限学习机对网络入侵进行分类,但该方法准确性较低。为此,提出一种基于特征选择的网络入侵检测模型研究方法。该方法首先以增强寻优性能为目标对网络入侵检测进行特征选择,结合分析出的特征选择利用特征属性的Fisher比构造出特征子集的评价函数,然后结合计算出的特征子集评价函数进行支持向量机完成对基于特征选择的网络入侵检测模型研究方法。仿真实验表明,利用支持向量机对网络入侵进行检测能有效地提高入侵检测的速度以及入侵检测的准确性。

The network intrusion model based on feature selection research

In order to effectively extract features from the malicious data collected to analyze, security network system security and stability, the need for network intrusion detection model is studied. But the current approach is to use genetic algorithm to find out the characteristics of the network intrusion subset of recycled for further selection of particle swarm optimization (pso), find out the optimal feature subset, finally using extreme learning machine classifying network intrusion, but this method has the problem of accuracy is low. Therefore, proposes a network intrusion detection methods based on feature selection. This method firstly in order to enhance optimal performance as the goal to feature selection of network intrusion detection, combined with analysis of characteristics of feature selection using the attributes of the Fisher than feature subset evaluation function is constructed, and combining with the feature subset of calculated results of evaluation function for support vector machine (SVM) to network intrusion detection based on feature selection methods. Simulation experiments show that support vector machine (SVM) is used to analyse the network intrusion detection can effectively improve the accuracy of the speed of intrusion detection and intrusion detection.