Machine Learning-Based Detection and Prevention of Anti-Money Laundering (AML) in the Financial Sector

Financial institutions, along with worldwideeconomic systems, face urgent threats from money laundering,which enables major illegal activities, including fraud andterrorism financing, as well as organized crime operations.Money laundering detection, alongside its prevention, requiressophisticated analytic systems to remain effective and efficient.Researchers have created a machine learning methodologybased on the Elliptic dataset to boost Anti-Money Laundering(AML) systems. The application of feature selection methodsserves to decrease data dimensions while improving systemperformance. XGBoost and Graph Neural Networks (GNN) andGraph Convolutional Networks (GCN) comprise the advancedclassifiers used for transaction classification. The evaluationtools used for model effectiveness measurement incorporate F1score and recall with accuracy as well as precision. Theexperimental results indicate that XGBoost outperforms GNNand GCN, achieving an accuracy of 97.5%, precision of 97.6%,recall of 97.4%, and an F1 score of 97.5%. These resultsdemonstrate that XGBoost is better than its competitors insuccessfully detecting fraudulent financial transactions whilebalancing recall and precision. A proposed frameworkpromotes efficient AML system development and enables moresecure financial environments built on trust. Future studies willexamine instant detection innovations and complex graphprocessing methods for combating new money launderingstrategies.