Proceedings of the 11th International Conference on Opportunities and Challenges in Management, Economics and Accounting
Year: 2024
DOI:
[PDF]
AI-assist auditing process a self-evolving model to capture ill behaviors
Jiarui Zeng
ABSTRACT:
This study employs advanced techniques to improve the performance of corporate fraud prediction. Following the conclusions drawn by Bao et al., we utilize raw accounting data instead of financial ratios, we adopt six AI models to forecast corporate fraud in the US. We extend a comprehensive set of fraud-related variables and organize them into three groups: Raw, Ratio, and Raw + Ratio. Among the six AI models tested, the Random Forest (RF) model outperforms the other five models in corporate fraud prediction. Additionally, we find that the model’s forecast is supplementary rather than conclusive.
keywords: fraud prediction, machine learning, random forest