Usage signatures analysis: an alternative method for preventing fraud in e-commerce applications

Abstract

From small local businesses to large multinationals, we can find a lot of initiatives in order to put their business online. E-Commerce is rapidly expanding worldwide with the emergence of new web shops, business to business services, or payment services, just to name a few. E-Commerce companies have been continuously improving their business incomes and extending their influence in the market. However, this proliferation of online business raised a very old phenomenon: fraud. Consequently, E-Commerce companies lose significant amounts of money every year due to a large diversity of fraudulent cases. This phenomenon is quite difficult to manage and handle, requiring significant efforts, both human and financial, in its identification and treatment. In this paper we present and discuss the use of signatures on the detection of potential fraud situations in an E-Commerce Web application, showing the way signatures are processed and analysed, as well as they can contribute to establish user behaviour deviation and consequently detect potential fraud situations in useful time.