Unraveling Patterns in Healthcare Fraud through Comprehensive Analysis
Khushi E Chirchi, B. Kavya · 2024
This study addresses healthcare provider fraud in Medicare, employing advanced machine learning models on a diverse dataset to predict potential fraud. The goal is to contribute insights for effective fraud detection and mitigate its impact on overall healthcare costs. Utilizing Logistic Regression, Random Forest, and addressing class imbalance through SMOTE, we discern patterns in provider behavior. Evaluation metrics, including confusion matrices, accuracy, sensitivity, specificity, Kappa values, AUC, and F1-scores, comprehensively assess each model’s performance. Findings highlight distinct behavior patterns and underscore SMOTE’s effectiveness in mitigating class imbalance challenges. Comparative analysis discusses algorithm strengths and weaknesses, offering insights for real-world implementation, impacting fraud prevention strategies for insurance companies, healthcare providers, and beneficiaries. In summary, this research makes a noteworthy contribution to the detection of healthcare fraud, offering insights for effective strategies. The incorporation of SMOTE enhances model robustness, leading to improved fraud detection capabilities and, consequently, reducing the impact of fraud on healthcare costs.