A Study on GA-Based WWN Intrusion Detection

Intrusion detection (ID) technology is a very important means to improve network security, therefore intrusion detection system should have a good performance. However, the traditional detection method cannot meet actual demands. Applying wavelet neural network (WNN) to intrusion detection has become a research focus recently, but some inherent defects in WNN reduce its efficiencies in operation and identification. Aiming at eliminating these limitations, we propose a new improved WNN algorithm basing on genetic algorithm (GA). This new algorithm combines WNN with GA in order to optimize network parameters and improve network performance. Making use of the actual data extracted from the intrusion detection system, this paper tests the newly constructed algorithm emulationally. The results show that used in intrusion detection, the new algorithm has higher identification rate and it can increase network running efficiency and the success rate of training. Thereby it can reduce undetected rate and false alarm rate. Thus it proves the validity of the new algorithm.