Title: Credit card fraud detection using moth-flame earth worm optimisation algorithm-based deep belief neural network

Authors: S. Deepika; S. Senthil

Addresses: Reva University, Rukmini Knowledge Park, Kattigenahalli, Yelahanka, Bengaluru, Karnataka 560064, India; Department of CSE, Anurag University, Hyderabad, India ' Reva University, Rukmini Knowledge Park, Kattigenahalli, Yelahanka, Bengaluru, Karnataka 560064, India

Abstract: Nowadays, credit card fraud actions happen commonly, which results in vast financial losses. Fraudulent transactions can take place in a variety of ways and can be put into various categories. Hence, financial institutions and banks put forward credit card fraud detection applications. To detect fraudulent activities, this paper proposes a credit card fraud detection system. The proposed system uses the database with the credit card transaction information and sends it to the pre-processing. The log transformation is applied over the database for data regulation in the pre-processing step. After, the appropriate features are selected by the information gain criterion, and the selected features are utilised to train the classifier. Here, a novel classifier, namely moth-flame earth worm optimisation-based deep belief network (MF-EWA-based DBN) is proposed for the fraud detection. The weights for the classifier are selected by the newly developed moth-flame earth worm optimisation algorithm (MF-EWA). The proposed classifier carries out the training and detects the fraud transactions in the database. The proposed MF-EWA-based DBN classifier has improved detection performance and outclassed other existing models with 85.89% accuracy.

Keywords: credit card transactions; fraud detection; information gain; earthworm optimisation algorithm; deep belief network.

DOI: 10.1504/IJESDF.2022.120021

International Journal of Electronic Security and Digital Forensics, 2022 Vol.14 No.1, pp.53 - 75

Received: 18 Apr 2020
Accepted: 31 Aug 2020
Published online: 04 Jan 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article