Design and development of adaptive authentication model to detect user behavior anomalies

Abstract

The password-based authentication system has recently become more secure as the riskbased authentication system(RBA) is indentured. Recent research in the area has shown the significant use of 2 Factor Authentication (2FA) and Multi-Factor Authentication(MFA) in many commercial applications using Risk Based newlineAuthentication(RBA). The RBA system monitors the parameters extracted during the user login process, and based on the proposed model, the system raises a multi-factor newlineauthentication to the user. As the vulnerability has increased concerning passwords, fingerprints easy access to any web application may result in a security flow; the reason can be the existing methodology of the RBA system and also the unavailability of the data of the users during the initial login process, which hinders the authentication system during the initial login process as there is no standard method to incorporate RBA in the authentication system. Few researchers have proposed novel approaches to improve the authentication system. Still, to the best of our knowledge, no research has suggested methods to address the authentication system during the initial login process and also provide a robust way, a combination of Machine Learning (ML) and statistical newlineapproaches. Hence, a novel method is proposed for the RBA system during the initial newlinelogin phase using a Hierarchical Sub-Feature Based Model -(HSFBM) for different user newlinecategories. The FAR is comparatively better in our proposed model against the standard newlinemodel, with minimal re-authentication requests for the user.