Leveraging Machine Learning for Improved Detection of Medicare Fraud
Velishala Aarthi, V.Sri Raghavendra, Vir Rao, Mrs.Hyma Birudaraju · INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT · 2025
Abstract— In order to overcome imbalanced datasets in healthcare fraud detection, the effort focuses on the Medicare Part B dataset. The hybrid resampling technique (SMOTE- ENN) is used with categorical feature extraction in this unique approach to balance the dataset. For fraud detection, logistic regression is used, and performance is assessed using a variety of measures. By using this method, problems with conventional resampling techniques like noise, overfitting, and information loss are lessened. The significance of AUPRC in situations with unbalanced data is emphasised by the study. Results demonstrate increased accuracy in detecting fraud, confirming the efficacy of the suggested approach.