The aim of the paper is the probability modelling of accesses to the different parts of e-learning course in interactive learning environment depending on time. This problem belongs to the open questions of the contemporary emerging discipline Educational Data Mining.We are concerned with the access probabilities to the individual parts of e-learning course content. For the purpose of modelling of the stakeholders’ behaviour dependence on time, we use multinominal logit model which is a special case of Generalized Linear Model. We pay attention to data preparation issues. Data about using e-learning courses are stored in databases. We assume that these data represent time data. Surprisingly, modelling of stakeholders’ behaviour dependence on time is missing in Educational DataMining discipline. The time variable, mostly stored in database in ”unixtime” form, integrating date and time, is partially used by sequence rules extraction, where it is only used for the tracking of visited e-learning course parts during each session. Therefore, we describe the applied multinominal logit model and methodology of the modelling in detail. We deal with parameter estimations. Finally, we figure that the multinomial logit model finds its application mainly in the process of restructuring the existing e-learning courses. We discuss about its possible contribution to the improvement of the interactive learning environment as well as in the personalization of the course content and structure.
Digital Object Identifier (DOI)
Munk, Michal and Drl?k, Martin
"Analysis of Stakeholders’ Behaviour Depending on Time in Virtual Learning Environment,"
Applied Mathematics & Information Sciences: Vol. 08
, Article 38.
Available at: https://dc.naturalspublishing.com/amis/vol08/iss2/38