Publicação

Money Laundering and Fraud detection under concept drift: A Systematic Review

Ver documento

Detalhes bibliográficos
Resumo:Financial Institutions are pivotal yet vulnerable agents in societies. Two of the threats they face are money laundering and fraud, requiring thorough risk mitigation strategies and quick response to avoid financial, operational and reputational costs. The evolving dynamics of these threats make it harder for traditional, rule-based detection systems to follow up, becoming obsolete over time. In contrast, machine learning’s ability to handle complex, nonlinear patterns and learn temporal sequences makes it a promising alternative. This study presents a systematic literature review that synthesizes a collection of peer-reviewed original works published between 2014 and 2024, identifying and categorizing machine learning approaches to each domain, with an emphasis on understanding how concept drift is handled. The analysis highlights prevailing algorithmic strategies, comparative insights between the two domains, knowledge gaps and future research directions to address them. The findings contribute to a structured understanding of machine learning applications in financial crime detection and strengthen decision-making towards more adaptive, data-driven detection systems.
Autores principais:Baptista, Simão Ortigão
Assunto:Money Laundering Fraud Financial Institution Machine Learning Concept Drift SDG 8 - Decent work and economic growth SDG 16 - Peace, justice and strong institutions
Ano:2025
País:Portugal
Tipo de documento:dissertação de mestrado
Tipo de acesso:acesso aberto
Instituição associada:Universidade Nova de Lisboa
Idioma:inglês
Origem:Repositório Institucional da UNL
Descrição
Resumo:Financial Institutions are pivotal yet vulnerable agents in societies. Two of the threats they face are money laundering and fraud, requiring thorough risk mitigation strategies and quick response to avoid financial, operational and reputational costs. The evolving dynamics of these threats make it harder for traditional, rule-based detection systems to follow up, becoming obsolete over time. In contrast, machine learning’s ability to handle complex, nonlinear patterns and learn temporal sequences makes it a promising alternative. This study presents a systematic literature review that synthesizes a collection of peer-reviewed original works published between 2014 and 2024, identifying and categorizing machine learning approaches to each domain, with an emphasis on understanding how concept drift is handled. The analysis highlights prevailing algorithmic strategies, comparative insights between the two domains, knowledge gaps and future research directions to address them. The findings contribute to a structured understanding of machine learning applications in financial crime detection and strengthen decision-making towards more adaptive, data-driven detection systems.