Information security risk assessment model based on optimized support vector machine with artificial fish swarm algorithm

Abstract

Because the information security risk assessment have the problem of less training data and slow convergence rate, we put forward a information security risk assessment model based on support vector machine (SVM) using artificial fish swarm algorithm (AFSA). In this paper, we used weekly security report of the government network security situation from China National Internet Emergency Center(CNCERT) as the source data [1]. We adopted the RBF function as the kernel function SVM, then optimized the penalty coefficient C and kernel function parameter 8 of artificial fish swarm algorithm. At the end of this paper, we established the optimal evaluation model for simulation. Our results showed that the information security risk assessment model based on AFSA SVM has higher accuracy and faster convergence rate than the one of cross-validation.