The support of e-learning platform management by the extraction of activity features and clustering based observation of users

Abstract

We present an application of data mining in e-learning, where web platform management was supported by the extraction of users' activity features and further by the clusterisation of users' profiles. By this approach we have identified groups of users with a similar activity on e-learning platform and were able to observe their performance. The experiments presented in this paper were performed on the real data coming from Moodle platform. Comparing to the other research in this filed, that focus on the analysis of students, we investigated teachers' behaviour. We have proposed a smoothing model in the form of a dynamic system, that was used to transform the logged events into time series of activities. These series were later used to cluster teachers' performance and to divide them into three groups: active, moderate and passive users. The main aim of our research was to propose and test an data mining based approach to support of e-learning management by an observation of teachers leading to an increase of the process quality. © Springer-Verlag Berlin Heidelberg 2010.