关键词: 入侵检测, 粗糙集理论, 核匹配追踪, 机器学习

Abstract: An intrusion detection method combinding Rough Set (RS) theory and Kernel Matching Pursuit (KMP) algorithm was proposed. In virtue of the ability that rough set can decrease the amount of data and get rid of redundancy, the proposed method could reduce the amount of machine training data, and overcome the disadvantage of slow processing speed because of large amount data in Support Vector Machine (SVM) algorithms. Meanwhile, KMP algorithm, with the help of good classification and recognition, made a different definition of importance on each sample, so that the core of property set could be classified, and intrusion detection could be realized rapidly with high classification accuracy and high detection rate and so on. The experimental results show that this method is better than the method based on RS only and the method based on Artificial Immune (AI).

Key words: intrusion detection, rough set theory, Kernel Matching Pursuit (KMP), machine learning