Credit card fraud (CCF) is a straightforward and appealing target. Online sites as well as E- commerce have enlarged their payment options, as risk of online fraud is increasing rapidly. Researchers began using various machine learning (ML) algorithms to detect and analyses online transaction fraud as fraud rates increased. This paper presents a random forest-based model to detect fraudulent transactions by analyzing customers' historical transaction details and extracting behavioral patterns. Cardholders are divided into groups based on the volume of their transactions. Then, using a sliding window method, the transactions performed by cardholders are aggregated from various categories. The behavioral patterns of the various groupings are then derived. The random forest (RF) classifier shows the greatest accuracy and hence proved to be one of the most excellent ways for detection/prediction of frauds. As a result, a feedback mechanism is implemented to address the issue of notion drift. The proposed model provides high accuracy of 99.99% with precision of 93% and recall of 73%. The proposed model provides better performance than isolation forest algorithm as well logistic regression and support vector machine.
Article Details
Unique Paper ID: 157570
Publication Volume & Issue: Volume 9, Issue 7
Page(s): 587 - 593
Article Preview & Download
Share This Article
Join our RMS
Conference Alert
NCSEM 2024
National Conference on Sustainable Engineering and Management - 2024