Fraudulent Detection in Healthcare Insurance

摘要

Our paper provides an extensive study of detecting fraudulent claims in healthcare insurance by leveraging machine learning algorithms. By using the publicly available medicare dataset, we are able to classify as fraud and non-fraud providers. Moreover, synthetically minority oversampling technique is used to avoid the class imbalance problem. Furthermore, a hybrid approach is used which is based on clustering and classification. Additionally, we have used other machine learning algorithms to check the efficiency of the best-suited algorithm.