Design of a risk based authentication system using machine learning techniques

Authentication provides a means to verify the legitimacy of a user trying to access any confidential or sensitive information. The need for protecting secure data hosted on the web has been rising exponentially as organizations are moving their applications online. Static methods of authentication cannot completely guarantee the genuineness of a user. This has led to the development of multi-factor authentication systems. Risk-based authentication, a form of multi factor authentication adapts itself according to the risk profile of the users. This paper puts forth the design of risk engine integrated with the system to examine the user's past login records and generate a suitable pattern using machine learning algorithms to calculate the risk level of the user. The risk level further decides the authentication method that the user will be challenged with. Thus the adaptive authentication model helps in providing a higher level of security to its users.