A Classification Algorithm of Online Network Traffic Based on Self-Organizing Incremental Radial Basis Network

An online semi-supervised traffic classification algorithm OSOINN-RBF is proposed to solve the problems of slow model training speed, high sample labeling cost and difficulty of real-time classification in traditional traffic classification methods. First, an improved self-organizing incremental neural network is used to perform online incremental unsupervised learning of data and to obtain a SOINN network that represents the distribution of input traffic data, and the weights and radius of the nodes in SOINN are used as the center and the radius of the hidden layer of the radial basis network; Then, the radial basis network captures the regularity that is difficult to find in the data, and has good generalization ability and learning convergence speed; Finally, a small number of labeled samples is used to adjust the weights of the output layer of the radial basis network and to improve the classification performance of the network. Experimental results and comparisons with mainstream classification algorithms show that the classification performance of the proposed algorithm is the best, with minimal time overhead; The classification accuracy of OSOINN-RBF algorithm is also improved by 5% to 7% compared with SOINN algorithm when facing unknown traffic categories. © 2020, Editorial Office of Journal of Xi'an Jiaotong University. All right reserved.