Abstract:

Due to the worse classification performance of classical SelfOrganizing Maps (SOM) for network traffic，a network traffic classification method based on the MIXKernel SelfOrganizing Maps(MIXKSOM) is proposed. Applying a mixed kernel function that is the linear combination of the radial basis function and the polynomial function to replace the Euclidean distance as distance measure, this method can not only combine the advantages of global and local kernel functions, but also simplify the complicated flow sample from the input space to the feature space. The experimental results show that this method has a better performance for classifying network traffic than the classification method based on the traditional SOM and the single kernel function SOM(KSOM), and get a better accuracy rate than NB.

Key words: traffic classification;selforganizing maps network;kernel function