An Fca Grounded Study of User Dynamics through Log Exploration

Nowadays the use of e-learning systems is increasing, especially in academia. Most researchers use Web Usage Mining (WUM) techniques to study the behavior of users of such systems. The results thus obtained are used to enhance e-learning systems or make different predictions. Current tools used for Web Usage Mining can offer a partial view on the usage of such systems. This paper presents the use of Formal Concept Analysis (FCA) as a Web Usage Mining instrument. It is also shown how different methods of FCA can be applied for revealing user behavior patterns within a system as well as users dynamics. The detection of users dynamics is a current research topic used to reveal how a system (e-learning instrument in our case) is built. Various authors who use FCA for Web Usage Mining propose ontology generation, semantic web personalisation, but none are using the whole range of FCA tools for exploration. In this paper, we debut by representing user dynamics patterns using ToscanaJ. Then, we use a triadic approach in order to investigate the behavioral patterns of students using the e-learning system. We also show how frequent triconcepts can be used for Web Usage Mining. Then, we use CIRCOS to represent the information content of the triadic data selection.