Using keystroke dynamics for gender identification in social network environment

Social networking is now a very widely adopted and highly pervasive communication medium, especially among younger people. However, while offering exceptional opportunities to share and interact these media also introduce the risk of transactions with individuals who deliberately conceal their identity or, importantly, can easily misrepresent their personal characteristics. This paper introduces an approach to addressing such risks by using a form of biometric data accessible from routine interaction mechanisms to predict important user characteristics, thereby directly increasing trust and reliability with respect to the claims made to message receivers by those who communicate with them.