Abstract
This paper uses machine learning (ML) and artificial intelligence (AI) algorithms to predict fraud in financial statements and contributes to an ongoing debate that uses ML and AI to address predictive-based problems in accounting research. The study uses the fraud triangle (FT) framework and the three strands of criminological theoretical propositions (strain, rational choice and neutralisation) to guide this inquiry. We used Random Forest, Classification and Regression Trees (CART), and Artificial Neural Network (ANN) Algorithms to test the theoretical propositions. We obtained the data from the U.S. Securities and Exchange Commission's (SEC) enforcement proceedings. The final sample consists of 104 most serious fraud cases heard by the SEC, which were coded for indicators representing the three legs of the FT: pressure, opportunity, and rationalisation. The results show that corporations are more likely to commit fraud when they experience the actual or perceived threat of financial strain. Corporations wi
Original language | English |
---|---|
Article number | 101441 |
Journal | British Accounting Review |
Volume | 56 |
Issue number | 6 |
Publication status | Published - 1 Nov 2024 |
Keywords
- Fraud, Machine Learning, Artificial Intelligence