Data Driven Detection of Trade Based Money Laundering (TBML): A predictive Analytics Framework for Securing US supply chains and Financial Integrity

Abstract

Trade-Based Money laundering (TBML) has become one of the most sophisticated and least detectable types of financial crime jeopardizing the integrity of the world trade and American financial system. The conventional detection tools have difficulty detecting advanced TBML cases because there is split information in the customs, banking, and trade finance sectors. This paper creates a predictive analytics account system which combines cross-sector data to improve detection of TBML and complements the U.S. Anti-Money laundering (AML) priorities. Based on a set of machine learning models such as supervised and unsupervised models and explainable artificial intelligence (XAI) algorithms such as SHAP and LIME, the presented framework can recognize the presence of hidden irregularities in trade, mispricing, and abnormal transactional practices, which are transparent and understandable. The framework proves to be much better in terms of predictive accuracy and interpretability, which allows the regulators and financial institutions to evaluate trade risks beforehand, without losing auditability and compliance. Combining the information on the customs declarations, financial transactions and finally the trade finance reports, the research offers a comprehensive view on the predictive trade risk analysis to enhance U.S. supply chain protection and financial integrity. The findings support the importance of explainable predictive analytics to make AML enforcement a transparent, data-based, and sustainable framework to protect national and international trade flows.