摘要： 进行入侵检测前必须分析输入数据的特征。使用粒子群优化算法对特征进行选择，消除冗余属性、降低问题规模、提高数据分类质量、加快数据处理速度。用二进制字符串序列表示粒子位置，阐述位置和速度的更新策略以及适应度函数的选择。在KDD CUP1999数据集上进行实验，结果表明与遗传进化算法相比，该方法可以更有效地精简特征，提高分类质量。

关键词: 粒子群优化算法, 特征选择, 入侵检测, 最优化

Abstract: It is necessary to analyze feature of input data before intrusion detection. This paper uses Particle Swarm Optimization(PSO) algorithm to select feature, eliminate the redundancy property, reduce the problem size, improve the quality of data classification and speed up the process. The position of the particle is expressed in a binary string. The update strategies of the position, velocity and the selection of fitness function are illustrated. The experiments with KDD CUP1999 and comparative results with Genetic Algorithm(GA) are described. It shows that the method is more efficient for feature selection and classification quality improvement.

Key words: Particle Swarm Optimization(PSO) algorithm, feature selection, intrusion detection, optimization

中图分类号: 

• TP309.2