Computer Network Security Evaluation Based on LM-BP Neural Network

Abstract

The assessment of computer network security is a significant system in computer network security assurance. Aiming at the shortcomings of BP neural network in network security evaluation, such as the slow convergence speed, the difficulty of globally optimal solution, the low accuracy of diagnosis and the uncertainty of network structure. Taking all these disadvantages into account, this paper aims to revise BP neural network by using levenberg-marquardt algorithm and combine the actual sample data to operate simulation test. As a result, LM-BP neural network algorithm, which possesses the advantages of fast learning speed and strong generalization ability, provides an effective, accurate and reliable method for the evaluation of computer network security.