Presenting the developed model of Benish model with emphasis on economic features using neural network, vector machine and random forest

Iranian Journal of Accounting, Auditing and Finance

مقالات آماده انتشار، پذیرفته شده، انتشار آنلاین از تاریخ 16 اردیبهشت 1401  

نوع مقاله: Original Article

شناسه دیجیتال (DOI): 10.22067/ijaaf.2022.75133.1103

نویسندگان

Kiumars Pourgadimi¹؛ Jamal Bahri Sales ¹؛ Saeed Jabbarzadeh Kangarloie¹؛ Akbar Zavar Rezaee²

¹Department of Accounting, Urmia Branch, Islamic Azad University, Urmia, Iran

²Department of accounting, Urmia University, Urmia, Iran

چکیده

As the business process becomes more complex, distortion risk of financial statements increases. In this regard, researchers have been looking for models to detect fraud in financial statements. Benish (1997) predicted earning manipulation using financial ratios and accruals. Since economic pressure is presented as external motivation of managers to manipulate income, in this study the Benish model is developed based on economic variables, which include Inflation Rate, GDP Growth, Exchange Rate and Economic Growth Rate. The fitting of the random forest, vector machine and neural network has been used to fit the extended model. The results show that accuracy of random forest model is 99.96% and more than the two neural network and vector model 96.1% and 93.62% respectively. In addition, developed model is more accurate than basic Benish model. The results show that economic factors plays great role in fraudulent financial reporting which should be considered when analyzing financial reporting.

کلیدواژه‌ها

Benish model؛ audit quality characteristics؛ neural network؛ vector machine and random forest