User Profiling Techniques: A comparative study in the context of e-commerce websites

Abstract

Today’s society is filled with a lot of information. It has become difficult to find the information we are looking for on the World Wide Web. When using e-commerce websites this is believed to lead to less sales. In order to facilitate this problem many websites can adjust the information that is presented to users based on the users’ needs. To achieve this several intelligent techniques can be used. In this study a selection of these techniques are explained and compared in order to give the audience a sense of the different possibilities accompanied by the strong aspects and pitfalls of these different techniques. The selected techniques are: Bayesian networks, decision trees, case-based reasoning, association rules and neural networks.