Ensemble Learning for Fraud Detection in E-commerce Transactions: A Comparative Study

Fraud detection is a critical measure in today’s digitalized world and e-commerce platforms. Considering the significant advantages that ensemble methods bring to the world of machine learning, it seems necessary to examine their potential usage in the field of fraud detection. This is especially important for e-commerce transactions where we need to assess whether Ensemble methods can be good candidates as effective classifiers. In this paper, we evaluate the potential of ensemble learning methods for fraud detection in e-commerce domain. We implement several well-known ensemble methods on an e-commerce customer data and compare their result using different performance criteria. Our results show that XGBoost and Random Forest outperform other ensemble methods for fraud detection. The results of this study can be helpful for those scholars who are willing to optimize their fraud detection systems with ensemble methods. Also, the present study shows which classification algorithms can be best used in an ensemble framework to be applied in fraud detection for online payments.

Ensemble Learning for Fraud Detection in E-commerce Transactions: A Comparative Study Keywords:

fraud detection , e-commerce , Ensemble Learning