AN IMPROVED NETWORK INTRUSION DETECTION METHOD BASED ON VQ-SVM

摘要

This paper proposes an improved efficient algorithm based on Vector Quantization and Support Vector Machine (VQ-SVM) for intrusion detection. The algorithm firstly reduces the network auditing dataset by using Vector Quantization technique, produces a training codebook, and then adopts fast training algorithm for Support Vector Machine to build intrusion detection model on the codebook. The experiment results indicate that the detecting efficiency of the intrusion detection model based on VQ-SVM algorithm can be greatly improved in comparison with the traditional SVM method, whereas the detecting accuracy doesn't decline somewhat.