Identification and Mitigation of Fraudulent Transaction using Deep Autoencoder

Authors

Vipin Khattri, Research Scholar
Department of Computer Science and Information Systems, Shri Ramswaroop Memorial University, Barabanki, India.
Sandeep Kumar Nayak, Associate Professor
Department of Computer Application, Integral University, Lucknow, India.

Abstract

In an ancient era, physical resources used to apply for transacting messages, treaty content, monarchy schemes, and policies and associated national or territorial currency which consumes time duration in the heavy count with negligible security. But as time passes, technological advancement has tendered its valuable and qualitative inputs to make the conventional transaction more better at its highest level of the extent, and as a qualitative and progressive resultant, the world is breathing in the current era of the digital environment with high-security priority. The responsibility of researchers and concerned authorities is to protect the online digital transaction under the safe digital environment. Therefore continuous enhancement is required in the upgrade of the security of the transaction system to handle digital transaction fraud. This research study suggests an approach of deep autoencoder for identifying fraudulent payment card transactions. To assess the outcome and validity of the projected approach of deep autoencoder for identifying fraudulent payment card transactions, testing was executed with the help of two datasets. The first dataset is a real credit card fraud dataset that is public available in world and the second dataset are generated by collecting the data using payment card transaction including genuine transaction and fraudulent transactions. A comparative analysis performed which is based on a comparison with different method and used first dataset. The proposed integration approach performed exceptionally with the different method and accomplished the maximum performance with respect to area under receiver operating characteristic curve (AUC) (95.66%).