Abstract: In order to improve the detection accuracy network intrusion detection, this paper proposes a novel network intrusion detection method, namely the BPSO-SVM-based detection algorithm that combines Binary Particle Swarm Optimization（BPSO） and Support Vector Machine（SVM）techniques to cope with feature selection issue for network intrusion. In the proposed algorithm, network intrusion detection is regarded as a multi-class categorization problem and feature subset is selected using a wrapper model, in which the BPSO searches the whole feature space and a SVM classifier serves as an evaluator for the goodness of the feature subset selected by the BPSO. The experimental results show that the proposed method reduces features dimensionality greatly and improves the detection accuracy of network intrusion as well as the significant improvement on detection speed.

Key words: network intrusion detection, binary particle swarm optimization, support vector machine, feature selection

摘要： 为了提高网络入侵的检测正确率，针对网络入侵检测中特征选择问题，将二值粒子群优化算法（BPSO）用于网络入侵特征选择，结合支持向量机（SVM）提出了一种基于BPSO-SVM的网络入侵检测算法。该算法将网络入侵检测转化为多分类问题，采用wrapper特征选择模型，以SVM为分类器，通过样本训练分类器，根据分类结果，利用BPSO算法在特征空间中进行全局搜索，选择最优特征集进行分类。实验结果表明，BPSO-SVM有效降低了特征维数，显著提高了网络入侵的检测正确率，还大大缩短了检测时间。

关键词: 网络入侵检测, 二值粒子群优化, 支持向量机, 特征选择