Concept drift and machine learning model for detecting fraudulent transactions in streaming environment

Abstrak

In a streaming environment, data is continuously generated and processed in an ongoing manner, and it is necessary to detect fraudulent transactions quickly to prevent significant financial losses. Hence, this paper proposes a machine learning-based approach for detecting fraudulent transactions in a streaming environment, with a focus on addressing concept drift. The approach utilizes the extreme gradient boosting (XGBoost) algorithm. Additionally, the approach employs four algorithms for detecting continuous stream drift. To evaluate the effectiveness of the approach, two datasets are used: a credit card dataset and a Twitter dataset containing financial fraud-related social media data. The approach is evaluated using cross-validation and the results demonstrate that it outperforms traditional machine learning models in terms of accuracy, precision, and recall, and is more robust to concept drift. The proposed approach can be utilized as a real-time fraud detection system in various industries, including finance, insurance, and e-commerce.