Application of Keystroke Dynamics Modelling Techniques to Strengthen the User Identification in the Context of E-commerce

Abstract

Keystroke dynamics is a biometric technique to identify users based on analysing habitual rhythm patterns in their typing behaviour. In e-commerce, this technique brings benefits to both security and the analysis of patterns of consumer behaviour. This paper focuses on analysing the keystroke dynamics against an e-commerce site for personal identification. This paper is an empirical reinforcement of previous works, with data extracted from realistic conditions that are of most interest for the practical application of modelling keystroke dynamics in free texts. It was a collaborative work with one of the leading e-commerce companies in Latin America. Experimental results showed that it was possible to identify typists with an accuracy of 89% from a sampling of 300 randomly selected users just by reading comment field keystrokes.