Behavioral Biometrics for Adaptive Authentication in Digital Banking - Guard Against Flawless Privacy

Abstract

This study throws light on the usage of Behavioral Biometrics in Digital banking arena. In current scenario more of physiological biometric modalities are being used in Digital Banking in comparison to psychological/ behavioral modalities which are difficult to record or mimic making behavioral biometrics more secure against replay attacks. This study suggests the usage of mobile screen swipe and touch data for user verification. The experiments were performed using publicly available UMDAA02 mobile swipe data set. The results of the study present a fine-tuned feature set for a swipe-based authentication system for mobile devices used for Digital Banking. It has been observed that with the suggested feature set, K-NN outperforms Naïve bays and SVM algorithms with the best EER of 14%. With the kind of performance shown by K-NN this study strongly suggests, swipe-based authentication system that can be used as a secondary layer of security in the digital banking system.