A New Method of Enhancing SOM''''s Abilities of Recognizing Consecutive Attacks
一种增强SOM网络识别连续型网络攻击的新方法

For the purpose of enhancing SOM's ability of identifying consecutive attacks, we present a FSOM model combing FIR filter and SOM, and give corresponding learning algorithm. To train the FSOM and test it's performance, the KDD benchmark dataset is employed. At last, the result is capable of detection rates of 93.1%, false positive rates of 7.3%, and shows the approach we used having a good performance in intrusion detection filed.