DDoS intrusion detection using Generalized Grey Self-organizing Maps

This paper describes the application of G2SOM (generalized grey self-organizing maps) to the DDoS(denial of service) intrusion detection. Generalized grey relational coefficients (G2RC), which characterize and stresses the whole correlation relationships between the input pattern and the weights of all the nodes that participate in competition, are explicitly introduced into the learning rule of the traditional SOM. In addition, SOM is generalized by the designed three G2RC functions, namely generalized grey self-organizing maps. Finally, the experiments on the DDOS datasets confirm their validities and feasibilities over the G2SOM in this paper. The dataset used is DARPA/KDD-99 publicly available dataset of features from network packets classified into normal and four DDoS attack categories.