A detection method for network security based on the combination of support vector machine

In this paper, we use a combination of support vector machine to improve the Standard SVM, which combine different kernel functions to improve the SVM' learning ability and generalization ability, thereby improving the performance of a combination SVM kernel function, and avoiding the assertiveness of the single prediction model. Combination forecasting model to make joint decisions on the results, making predictions more accurate. At the same time, taking advantage of Particle swarm optimization algorithm to overcome existing problems: the poor randomness and global of support vector machine parameters. And the particle swarm because of its global search capability, simple model, fast convergence, has a great advantage in dealing with the problem of high dimension, and the Particle swarm optimization in this paper is an improved Particle swarm optimization algorithm.