Abstract
Background: The increase in financial fraud has become a significant global challenge, threatening institutions, stability, and personal security. The adoption of advanced fraud detection models has increased significantly, especially those that leverage anomaly detection to identify unexpected patterns in financial transactions. Whereas other various approaches have been proposed, including machine learning and artificial intelligence techniques, limited studies have thoroughly documented the research landscape in this domain.
Objectives: Determine the most prominent fraud-detection models in the financial sector for anomaly recognition.
Method: A bibliometric analysis was conducted using publications from the Web of Science database, encompassing articles published between 2006 and 2025. The analysis used performance and scientific mapping approaches, such as keyword co-occurrence and VOSviewer citation network visualisation, to evaluate publication trends and influential authors.