The Role of Machine Learning in Detecting Earnings Manipulation

Abstract

Earnings manipulation poses a persistent threat to the integrity of financial reporting, undermining investor confidence and market stability. Traditional detection methods, while useful, often fail to identify increasingly sophisticated manipulative practices. This study investigates the role of machine learning (ML) in detecting earnings manipulation by employing a mixed-methods approach that integrates quantitative modeling and qualitative interpretation. Using a dataset comprising financial statements, governance indicators, and enforcement records, we develop and evaluate multiple ML models including logistic regression, random forest, gradient boosting, support vector machines, and deep neural networks. The results demonstrate that ensemble-based algorithms, particularly gradient boosting and random forest, consistently outperform traditional methods in terms of accuracy, recall, and F1-scores. Deep neural networks further enhance detection by capturing non-linear patterns, though their interpretability requires augmentation through explainable AI (XAI) frameworks such as SHAP and LIME.The experimental findings, supported by nine comprehensive tables and twelve figures, confirm that governance-related variables, accrual-based indicators, and volatility-sensitive features significantly improve predictive power. Moreover, out-of-sample testing highlights the scalability and robustness of ML models across industries and geographic contexts. The integration of XAI ensures that predictions are transparent and interpretable, enabling auditors and regulators to understand the drivers of manipulation risk. Collectively, the results suggest that ML not only enhances the accuracy of fraud detection but also offers proactive monitoring capabilities, thereby shifting detection from retrospective auditing to preventive financial governance. The study contributes to the literature on AI in accounting and provides practical implications for regulators, policymakers, and auditing professionals aiming to strengthen corporate accountability.