Extreme Gradient Boost Classifier based Credit Card Fraud Detection Model
Deepak Singh Nijwala, Sudhanshu Maurya, M. P. Thapliyal, Rohan Verma
Abstract
Financial fraud has evolved into an extremely serious problem in the business world. The effects of financial fraud on businesses, the economy, and people’s standard of living are widespread. Credit card fraud describes the shady practice of using personal information for online purchases. As the use of credit cards grows, so does the prevalence of fraud involving these payments. These days, credit card transactions are commonplace not only for online but also for in-store purchases. However, it has also made it easier for scammers to take advantage of this great opportunity. We can identify fraudulent transactions using Credit Card Fraud Detection. To handle imbalanced data, the proposed model in this work employs the XGBoost classifier to detect fraud transactions. The typical technique pre-determines the threshold value, resulting in inefficiency. Thus, in our proposed approach, several threshold values are computed and compared to identify the ideal value that provides an optimal outcome and high efficiency.