Automating semantic annotation to enable learning content adaptation

Abstract

This paper presents an approach to automatic annotation of learning objects' (LOs) content units that can be later assembled into new LOs personalized to the users' knowledge, preferences, and learning styles. Relying on a LO content structure ontology and some simple content-mining algorithms and heuristics, we manage to rather successfully determine the values of metadata elements aimed at annotating content units. Specifically, in this paper we present the specificities of generating metadata that describe the subject (based on a domain ontology) and the pedagogical role (based on an ontology of pedagogical roles) of a content unit. To test our approach we developed TANGRAM, an adaptive web-based educational environment for the domain of Intelligent Information system that enables on-the-fly assembly of personalized learning content out of existing content units.