Financial Statement Fraud Risk Analysis Using Fuzzy Logic

As financial markets are characterized by the increasing incidence and complexity of fraudulent activities, generating substantial losses for companies while posing serious risks to potential investors, financial fraud is a constantly debated issue, particularly with regard to its prevention and detection. As classical methods for detecting financial fraud have proven inefficient and time-consuming, many researchers have turned to artificial-intelligence-based methods. The use of AI-based methods for fraud detection is currently a widely debated topic, particularly regarding fraudulent financial statements. This study aims to determine the extent to which fuzzy logic contributes to financial statement fraud risk assessment. The target population comprises Romanian companies listed on the Bucharest Stock Exchange. Following analysis, a sample of 62 listed companies was selected. The analysis covers the last seven completed financial years (2018–2024). The dependent variable is represented by financial statement fraud risk (FSF), measured using the modified F-score model, under the influence of some independent variables that are defined by a series of financial ratios: return on assets, return on equity, net profit margin, leverage and working capital. Additionally, taking into account the audit opinion type and advanced statistical methods for data analysis, the research results demonstrated the importance of using fuzzy logic in fraud risk assessment by improving the process of detecting fraud in financial statements, based on specific financial ratios and statistical associations between these ratios, as fuzzy rules.