Research on the national security risk assessment model: a case study of political security in China

National security risk assessment is an indispensable key component of the national security system, serving as a foundation for national security decision-making, crisis prevention, and protection of national interests. While national security encompasses a broad spectrum of risks, including political, military, economic, technological, and social dimensions, this study focuses specifically on political security risk assessment as a case study within the broader framework of national security. Currently, there is no consensus on standardized requirements, processes, and methods for assessing national security risks. This study aims to develop a comprehensive national security risk assessment model, and the assessment path is explored from three dimensions: risk control processes, relationships among assessment factors, and the assessment model itself. To validate the model’s feasibility, we conduct an empirical study focusing specifically on China’s political security as one critical domain within the broader national security landscape. Using a game theory combination weighting method, risk indicator weights were determined, and political security risk assessment was calculated through the simulated annealing optimized projection pursuit evaluation method. The results indicate that China’s political security risk has generally declined, dropping from 76.39 in 2001 to 24.99 in 2022, with projections suggesting a continued decline to 11.00 by 2032. Among the five predictive models compared, the BP neural network model achieved the highest determination coefficient (R2 = 0.9967), demonstrating superior predictive performance. The proposed model provides a methodological framework that can be applied not only to political security but also adapted to other domains of national security, offering theoretical guidance and methodological support for comprehensive national security risk evaluation.