摘要：The aim of this study is to determine the fraud risk of the independently audited financial statements of companies whose stocks are traded in the Borsa Istanbul Textile, Clothing and Leather sector between 2015-2019 using data mining-based methods, through financial ratios, and accordingly, to reveal the success of these methods in detecting fraud. For this purpose, independent audit reports and weekly Capital Market Boards of Turkey (CMB) Bulletins were examined within the scope of the study, and cases of applying to fraudulent financial reporting practices were determined. In this context, 127 financial statements and independent audit reports of relevant periods were examined. In the study, 12 financial ratios used in literature to explain fraudulent financial reporting and 10 methods based on data mining were used. According to research findings, all models based on data mining used within the scope of the study were more than 70% successful in correctly classifying financial statements that are considered to have fraud risk and financial statements that are considered to have no fraud risk, and the most successful methods are models established with J48 and Deep Learning methods.
