Imbalanced Classification Problems: Systematic Study and Challenges in Healthcare Insurance Fraud Detection

The recent developments made in the data mining technologies have greatly influenced the data classification process. The growth of applications has increased the volume of the data and thus, the classification task becomes quite complex. Due to the uncertainties and unbounded nature of the data, class imbalance is one of the significant issues which determine the performance of the classifiers. In this paper, we present the challenges of the imbalanced classifications in the healthcare insurance claiming frauds. Most classification algorithms make use of majority class by ignoring the minority class. There are different approaches available to deal with the imbalance datasets which is reviewed in this study. A systematic study is done on each approach which presents the challenges pertaining in the class imbalance issues.