Bibliometric analysis of trends and directions in accounting fraud

Accounting fraud is a pervasive problem that presents significant challenges for financial reporting, corporate governance and investor protection. It can lead to substantial financial losses and damage the reputation of companies and auditors, which is why it is an important topic of research in accounting and finance literature. This study proposes a framework for integrating machine learning technologies into the analysis of accounting fraud literature. This framework allows the examination of a significant amount of literature to highlight current trends in the field and identify potential themes for future research. The originality of this paper lies in the adoption of machine learning technologies to analyze the literature devoted to accounting fraud, an approach that distinguishes this study from traditional literature review methods. Through the specific application of Latent Dirichlet Allocation (LDA) thematic modeling, our work goes beyond conventional approaches, allowing the automatic detection of primary themes and topics across a large volume of scholarly work.