Artificial Intelligence and Intelligent Analytical Models in the Evaluation of Audit Report Quality: Evidence from the Tehran Stock Exchange

Abstract

Audit report quality is one of the central mechanisms through which public trust in financial reporting and information transparency in capital markets are strengthened. With the expansion of artificial intelligence and analytical technologies in accounting and auditing, intelligent models provide new opportunities for evaluating audit report quality through data-driven prediction. This study aimed to design and compare intelligent analytical models for evaluating audit report quality among companies listed on the Tehran Stock Exchange. The study used a quantitative empirical design. Audit report quality was operationalized as the issuance of a modified audit opinion. The independent variables included financial leverage, firm size, profitability, auditor tenure, audit fee, operating cash flow ratio. The sample consisted of 150 listed companies over the period 2017–2023. Logistic regression, decision tree, random forest, and gradient boosting models were used to analyze the data. Model performance was evaluated using accuracy, precision, recall, F1-score, and ROC-AUC. The results indicated that financial leverage had a positive and significant association with the probability of receiving a modified audit opinion, whereas firm size, profitability, auditor tenure, and operating cash flow were negatively associated with modified audit opinions. Among the predictive models, random forest achieved the strongest performance (accuracy = 0.83, precision = 0.81, recall = 0.79, and ROC-AUC = 0.88). Variable-importance analysis showed that financial leverage, firm size, auditor tenure, and profitability were the most influential predictors. The findings suggest that intelligent analytical models, particularly random forest, can support audit-risk assessment, regulatory monitoring, and investor decision-making in the Iranian capital market.