An isolation forest based suspicious transaction detection model for Anti money laundering. A

Abstract

Abstract In response to the evolving financial landscape and the persistent challenge of preventing illicit activities, this thesis introduces an innovative Isolation Forest-based model, named iForest, for identifying suspicious transactions with a specific focus on countering money laundering. Rooted in practical considerations and a nuanced understanding of the challenges faced by the banking and financial sectors in Sri Lanka, the research delves into effective methods for detection. The iForest model leverages machine learning techniques to enhance the accuracy and efficiency of detecting anomalous financial activities, offering a significant improvement over traditional method. The thesis encapsulates comprehensive insights, methodologies, and findings aimed at advancing the current understanding and capabilities in addressing the critical issue of money laundering. Through empirical analysis and real-world data application, the research demonstrates the model's robustness and adaptability to various financial environments. Key contributions include the model's scalability, real-time monitoring capabilities, and integration with existing AML frameworks, making it a practical tool for financial institutions. Furthermore, the thesis discusses the ethical implications and compliance aspects of deploying such advanced technologies, ensuring that the iForest model aligns with regulatory standards and respects individual privacy. By providing a detailed evaluation of the model's performance and its potential impact on the financial sector, this research offers a valuable resource for policymakers and practitioners seeking to enhance AML efforts and protect financial systems from illicit activities. Keywords - Anti-Money Laundering, Isolation Forest, Financial Crime Detection, Machine Learning Models, Anomaly Detection, Regulatory Compliance, Semi-Supervised Learning & Data pre-process