Learning user profiles to model adaptive interfaces

Abstract

The primary objective of this research is to offer concepts related to the “user interface”, as part of a knowledge-based system whose function relations with the user. We believe it is imperative to address the problem of user interface from an ergonomic point of view taking advantage of methods and tools that have been developed in this regard. One issue that is of real interest in the mining area of use of the web is to capture user activities during connection and extract patterns of behavior to define their profile in order to design adaptive interfaces. Second, the context-free grammars probabilistic modeling that allows browsing sessions is used. That renders Web sessions by graphs and probabilistic context-free grammars so that the sessions that are most likely are considered the most popular or most preferred, therefore the most important in relation to a particular topic. It aims to develop a tool for processing Web meetings obtained from server log represented by probabilistic free grammars of context.