Document Type : Original Article
Authors
1
Professor of Accounting Department, Faculty of Economic and Administrative Sciences, Mazandaran University, Babolsar, Iran
2
PhD student in accounting, Faculty of Economic and Administrative Sciences, Mazandaran University, Babolsar, Iran
3
PhD student in accounting, Department of Accounting, Central Tehran Branch, Islamic Azad University, Tehran, Iran
4
Graduated with a doctorate in accounting, Faculty of Accounting and Financial Sciences, University of Tehran, Iran
10.22034/aafie.2024.413225.1033
Abstract
Today, knowledge is considered as a valuable and strategic resource and an asset for evaluation and prediction, and it leads to providing solutions in the field of detecting fraudu in the financial statements of companies, which increases accuracy and reduces ineffective labor to investigate and detect fraudulent companies. By using a solution like the proposed solution, it is possible to investigate and discover frau in the financial statements of companies full-time, and this does not require human labor, but the system itself can intelligently perform the diagnosis and inform. In the past, various solutions for detecting fraud were presented, each of which had problems. Therefore, the present research presents a method to detect fraud in the financial statements of companies with the help of artificial intelligence methods including machine learning algorithms. For this purpose, at first, after data preprocessing and data transfer, features (independent variables) were selected using the combined algorithms of the Ruff set and hierarchical analysis, and by training, calculating and testing the weights of these features through the algorithm Machine learning models of these algorithms were presented to predict the fraud of financial statements. Finally, the prediction accuracy of the proposed method was checked with some of the previous methods and the results indicate that the proposed method performs better than them.
Keywords
