Integrating Machine Learning and Rule-Based Systems for Fraud Detection - A case study based on the Logic Learning Machine

Aug 5, 2026

Money laundering presents a major global risk with far-reaching effects on economies and international security. Detecting suspicious transactions remains difficult because criminal strategies constantly change and analysts must process massive daily data volumes. This study introduces a hybrid approach combining machine learning with heuristic rules to improve identification of fraudulent activity. Using the SAML dataset of millions of bank transactions, which is highly imbalanced, the method evaluates heuristic rules via covering and error metrics and integrates them into the Logic Learning Machine for performance comparison.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages24

Reading Time1 h 0 min

AccessFree to download

Introduction

Money laundering context and impact

AML objectives and operational workflow

Challenges in AML detection

RISK MANAGEMENT MAGAZINE Vol. 20, Issue 2 May – August 2025 EXCERPT [https://www.aifirm.it/rivista/progetto-editoriale/](https://www.aifirm.it/rivista/progetto-editoriale/) Integrating Machine Learning and RuleBased Systems for Fraud Detection: A case study based on the Logic Learning Machine Pier Giuseppe Giribone, Giorgio Mantero, Marco Muselli and Damiano Verda Integrating Machine Learning and Rule-Based Systems for Fraud Detection: A case study based on the Logic Learning Machine Pier Giuseppe Giribone (University of Genoa, BPER Group); Giorgio Mantero (Rulex Inc.); Marco Muselli (Rulex Inc.), Damiano Verda (Rulex Inc.) Corresponding Author: Giorgio Mantero ([giorgio.mantero@rulex.ai](giorgio.mantero@rulex.ai)) Article submitted to double-blind peer review, received on 14st June 2025 and accepted on 13th August 2025 Abstract Money laundering is one of the most relevant global challenges, with significant repercussions on the economy and international security. Identifying suspicious transactions is a key element in the fight against the phenomenon, but the task is extremely complex due to the constant evolution of the strategies adopted by criminals and the great amount of data to be analyzed daily. This study proposes a hybrid method that integrates Machine Learning models with heuristic rules, with the aim of identifying fraudulent transactions more effectively. The dataset used, SAML, includes millions of bank transactions and presents a strong imbalance between classes (fraudulent vs regular transactions) . The entire process was carried out through a self-code platform designed to optimize data management, processing and analysis. The heuristic rules were evaluated using the covering and error metrics and then integrated into the Logic Learning Machine (LLM) task. The effectiveness of the approach was verified by comparing two main configurations: one based exclusively on the use of LLM and the other combining LLM and heuristic rules. The results obtained highlight that the integration of heuristic rules improves the performance of the model, confirming the synergy between Machine Learning and expert knowledge. This study confirms the effectiveness of the hybrid approach and emphasizes the importance of the union between automated analysis and human insight to address the challenges posed by money laundering. Key Words: Anti-Money Laundering (AML), Transaction Monitoring, Synthetic Dataset, Machine Learning (ML), Heuristic Rules, Logic Learning Machine (LLM) JEL code: C38, C45, K14, K22 1) Introduction As described by the United Nations Office on Drugs and Crime (UNODC): “Money laundering is the processing of criminal proceeds to disguise their illegal origin. This process is of critical importance, as it enables the criminal to enjoy these profits without jeopardizing their source” (UNODC, 2021) . Money laundering therefore indicates all those processes implemented by criminal organizations in order to disguise the origins of money obtained through illegal activities, such as corruption, drug trafficking, or fraud, to make it appear legitimate. By implementing it, such organizations can integrate illicit funds into the financial system, allowing further investment in illegal operations while still managing to avoid recognition by supervisory bodies. These illicit activities pose serious concerns for the global economy because they are used to fund criminal activities, they disrupt financial markets and ultimately may also damage the reputation of the financial institutions involved in fraudulent activities. Although it is clearly not possible to directly measure the extent of money laundering as we usually do with legitimate economic activities, its scale is massive, thus representing a significant threat to global financial systems. The UNODC estimates that money laundering accounts for 2-5% of global GDP annually, i.e. between 800 billion and 2 trillion EUR (UNODC, 2021), thus unders

FAQs

Why is fraud and money-laundering detection so challenging in practice?

Criminal strategies evolve continuously and vast amounts of data must be analyzed daily, making suspicious pattern identification highly complex.

What is the core idea of the proposed method in the study?

The study proposes a hybrid approach that integrates machine learning models with heuristic rules to identify fraudulent transactions more effectively.

How were the heuristic rules assessed and incorporated into the Logic Learning Machine task?

Heuristic rules were evaluated using covering and error metrics, and then integrated into the Logic Learning Machine (LLM) for the final comparative configurations.

Aug 5, 2026

Upload by Evangeline