فهرست نشریات علمی دسترسی آزاد ایران

This research aims to design a data analysis-based learning model for training in accounting fraud detection in digital environments. The research was conducted in an exploratory and survey manner. The qualitative part was conducted using the Delphi method and the quantitative part using the interpretive structural model and structural equations method. In the qualitative part, the population studied was managers and experts in financial management and information technology management, 12 people were selected using the available judgment sampling method. Using the library technique, the components affecting the model were identified. First, the identified components were screened and evaluated using the Delphi technique. Then, modeling was carried out using the Delphi method and the interpretive structural method. The software used were EXCEL and MICMAC. Using the Cochran formula, 384 people were selected. Based on a researcher-made questionnaire based on qualitative analysis, data analysis was performed using coding and path analysis. Data analysis was performed using coding and path analysis. Based on sampling, the structural equation modeling technique in SMARTPLS software was used to fit the proposed model. The ten main criteria, including big data analytics, fraud machine learning, audit artificial intelligence, smart transaction tracking, encryption and transparency, financial process automation, hidden behavioral data mining, digital anomaly detection, financial blockchain platform, and continuous real-time monitoring, are classified into seven levels. This hierarchical classification shows that some criteria play a fundamental and fundamental role in the success of fraud detection systems, while others affect the operational, analytical, and monitoring layers more.