关键词: 入侵检测, 信息安全, 主成分分析

Abstract: With the rapid development of high-speed network，how to capture the unusual characteristics of the attack quickly and effectively in high-speed network has become an important problem of the IDS.Using the character of principal component analysis of the various principal component is not relevant and the principal component is linear combination of all the original features，the high-dimensional feature vectors are effectively mapped to the low-dimensional space.It has not only maintained the characteristics of the original data，but also reduced the discarded package rates under the high-speed network system.Through testing the experiment of KDD Cup99 data and using of BP neural network classifiers are proved that the method is correct and effective.Meanwhile，the data management capabilities module is designed，it makes the algorithm and the practical application more closely and improves the intrusion detection system’s overall performance.

Key words: intrusion detection, information security, principal component analysis