A security risk assessment approach based on improved PSO-BPNN for intelligent network information systems

Smart cities have brought numerous benefits to improve citizens’ living quality, but they have also brought greater cyber security challenges. Risk assessment is a commonly used approach to enhance the cyber security of information systems. However, information systems in smart cities are usually large-scale and highly dynamic intelligent network information systems where traditional risk assessment approaches are not applicable. To address this problem, this paper applies artificial intelligence (AI) technology to perform risk assessment as AI-based approach can process big security datasets with fast classification speed and has the ability to discover hidden patterns with higher accuracy. We improved the particle swarm optimization (PSO) algorithm by constructing space symmetric particles and dividing the search process of the PSO algorithm into two stages. On this basis, an IPSO-BPNN (Improved Particle Swarm Optimization-Back Propagation Neural Network) algorithm for security risk assessment of intelligent network information systems is proposed. Following ISO/IEC 27005 and ISO/IEC TS 5689 standards, a risk assessment criterion and risk factors for intelligent network information systems are presented. Also, the overall process of the proposed risk assessment method and the calculation method of risk characteristics are presented. We evaluated the proposed method by comparing it with existing novel AI-based cybersecurity risk assessment methods. Experimental results show that the proposed method can accurately assess cyber security risks in intelligent network information systems with high efficiency.