Abstract The use of Social Network Analysis (SNA) for online learning communities’ analysis is common and usually performed after the ending of semester. Yet, even if such analysis is very useful, it is costly, and cannot be performed many times during the semester. In this paper, we present a model of automated SNA based inference, for a large- scale community, taking into account specific environment of developing higher education system. The model is designed so to send automated reminders to all users, according their activity in the period of two weeks. One additional analysis after the mid-term exams checks if activity matches performance. It has crucial role in directing both students and educators towards the common goal: success at the final exams. The presented model enables inference on user attributes, which are stored in student model ontology. As such, the model is a step in the development of semantic and adaptive learning environment.

Abstract FIT has a large virtual community, developed for DL students, but including the growing number of in-situ students. Through the last five years, most of the students have accepted and used FIT Community Server (FITCS) for interactions, but the extent of knowledge sharing was never measured. It was determined that FITCS is a small-world scale-free network, with several hubs: some administrators and some spammers. Now we determined characteristics of the social network, such as density, centrality, degree, closeness and betweenness, for overall communication in first two semesters and for two subjects in order to estimate knowledge sharing.