Entropy Based Support Vectors And Ant Colony Networks For Intrusion Detection

摘要

Intrusion detection normally generates a large number of false rate alerts particularly in large scale networks. That results in the dispute on efficiency and accuracy of network attack detection. In this paper for the classification of data as normal and abnormal, a new algorithm ESVAC (entropy based support vectors and ant colony networks) is proposed that is employed to network intrusion detection to take rewards of two algorithms keeping off their disadvantages. The algorithm followed out and valuated using KDD99 data set. As a result, experiments show that ESVAC performs better in price of training time and detection accuracy.