Abstract

Objective: In today's business world, many companies are not safe from financial abuse. Financial fraud can cause financial losses and reduce the reputation of companies and reduce customer trust. This global trend has forced everyone to look for ways to prevent fraud. To this end, the purpose of the present study is to identify the factors affecting internal audit performance using new deep learning techniques to rank the factors for detecting financial statement fraud.

Method: The spatial scope of this study includes companies traded on the Tehran Stock Exchange and its temporal scope includes the years 1391 to 1400 (solar calendar). Library methods were used to collect the data required for this study. In the section examining the factors affecting internal audit performance, the regression analysis method was used. Then, in the next section, deep learning methods and artificial neural networks were used for the study.

Findings: The results showed that, compared to the regression results, deep learning models and artificial neural networks performed better for the justification parameter in the coefficient of determination index and mean square error. Also, the findings showed that the performance of deep learning models was better than that of artificial neural networks.

Conclusion: The use of deep learning techniques can be valuable and effective for identifying factors affecting internal audit performance and ranking factors for detecting financial statement fraud.