Healthcare Insurance Fraud Prediction with Correlation based Forward Feature Selection
Sushma Rath, Suvasini Panigrahi · 2023
Health insurance serves as a vital safety net for individuals of various economic backgrounds in contemporary society. Enormous financial losses and loss of life have occurred throughout history due to healthcare fraud. The primary motive for fraudsters engaging in healthcare fraud is financial gain. Fraudsters can be doctors, patients, insurance companies, or combinations of these entities, including collaborations between doctors and patients, patients and insurance companies, and more. In this work, various phases of machine learning pipeline are executed, including collecting data, preparing it, building models, and evaluating their performance. The paper presents a novel combination of correlation-based and forward feature selection based on feature importance techniques to enhance overall performance.In this study, a comprehensive comparison of various supervised learning algorithms is conducted, namely Logistic Regression (LR), Decision Tree Classifier (DT), Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Network (ANN). The Random Forest approach yielded the highest accuracy, achieving a remarkable mean accuracy rate of 93.83%.