基于量子进化算法的网络入侵检测特征选择

针对当前网络入侵检测中普遍存在检测速度较慢的缺陷，提出了一种新的网络入侵检测特征选择方法。该方法将量子进化算法应用于网络入侵检测的特征选择，从网络连接的原始特征属性中选出一组有效的特征用于入侵检测，以提高检测效率。首先以增强寻优性能为目标改进了量子进化算法，基于特征属性的Fisher比构造了特征子集的评价函数，然后按照量子进化算法的流程设计了网络入侵检测特征选择算法。通过KDD99样本数据集的实验，表明算法是有效的，既保证了入侵检测的分类性能，也提高了入侵检测的效率。

Feature selection for network intrusion detection based on quantum evolutionary algorithm

Concerning the disadvantages of slow detection speed in current network intrusion detection, a new feature selection method of network intrusion detection was put forward. The method applied Quantum Evolutionary Algorithm (QEA) to feature selection of network intrusion detection, extracted an optimal subset used in intrusion detection from the original feature set in network connections, so as to get better detection efficiency. First, QEA was improved in order to make its searching performance better, and the criterion function of feature subset was constructed based on the Fisher ratio of feature attributes. Then, the feature selection algorithm of network intrusion detection was designed according to QEA flow. Last, experiments were carried out using the sample data from KDD99. The experimental results show that the proposed algorithm is effective, and it can not only ensure the classification performance of intrusion detection but also improve the detection efficiency.