International Risk Early-Warning System Construction and Application for International Pharmaceutical and Chemical Trade: An Empirical Analysis Based on Multi-Source Data Fusion and LSTM-PSO Algorithm

Abstract

International pharmaceutical and chemical cross-border trade is characterized by high compliance requirements, significant market volatility, and long supply chain links. Small and medium-sized foreign trade enterprises (SMEs) generally face the pain points of unpredictable compliance risks, lagging market responses, and insufficient supply chain early warnings. To address these issues, this study focuses on four core risks: compliance, market, supply chain, and credit. A four-dimensional risk assessment framework comprising 23 indicators is constructed. By integrating multi-source data from policies, markets, and enterprises, and optimizing the key parameters of the Long Short-Term Memory (LSTM) network using the Particle Swarm Optimization (PSO) algorithm, an intelligent risk early-warning system based on the LSTM-PSO algorithm is established. An empirical analysis is conducted using the cross-border trade data of Wuhan Kuda Hui Trading Co., Ltd. from 2019 to 2024. The results show that the model achieves a high-risk event early warning accuracy rate of 92.3%, and the lead time for logistics delay risk early warning is extended to 168 hours. After the system is implemented, the incidence of high-risk events in the enterprise decreases from 15.6% to 4.8% (Benamor, W. D., 2022), and in 2024, a loss of 2.16 million yuan is avoided. Moreover, the relevant solutions have been promoted to 12 enterprises in the industry, with an average reduction in risk losses of 37.2%. The study confirms that the LSTM-PSO risk early-warning system can effectively enhance the risk prevention and control capabilities of SMEs in the pharmaceutical and chemical foreign trade sector and has significant practical application and industry promotion value.