A Wireless Intrusion Detection Method Based on Dynamic Growing Neural Network

In this paper an intrusion detection method based on dynamic growing neural network (DGNN) for wireless networking is presented. DGNN is based on the Hebbian learning rule and adds new neurons under certain conditions. When DGNN performs supervised learning, resonance will happen if the winner can´t match the training example; this rule combines the ART/ARTMAP neural network and WTA learning rule. When DGNN performs unsupervised learning, post-prune is carried out to prevent overfitting the training data just like decision tree learning. The intrusion detection method is an anomaly detection method and the feature is selected from the packets. In the experiments, we first check the ability of the neural network and then use it to perform detection in a WLAN. The results show that it can detect new intrusion behavior and some improving methods are presented in the conclusions