Predictive Analytics and Artificial Intelligence for Regulatory (RegTech) Compliance in the Financial Industry

Background: The increasing prevalence of unethical practices in the financial sector has led to stricter regulatory frameworks, driving the demand for advanced Regulatory Technology (RegTech) solutions. Machine learning offers a promising approach to enhance compliance monitoring. Methods: This study applies RandomForest (RF), Gradient Boosting (GB), and AdaBoost (AB) to predict compliance risks using the bank marketing dataset. SMOTE and cost-sensitive learning addressed class imbalance, while Grid Search Cross-Validation optimized model performance. Accuracy, precision, recall, and F1-score were utilized for evaluation. Results: RF achieved the best performance with 93.98% accuracy, 92.03% precision, 96.42% recall, and a 94.17% F1-score, followed by GB and AB with slightly lower results. Traditional models like K-Nearest Neighbors (KNN) and Extreme Gradient Boosting (XGB) performed less effectively. Conclusion: The research shows that financial institutions may get a better handle on regulatory conformity with the use of RegTech solutions based on ML, which greatly enhance compliance monitoring.