Academic journal article Educational Technology & Society

Semantic Linking of Learning Object Repositories to DBpedia

Academic journal article Educational Technology & Society

Semantic Linking of Learning Object Repositories to DBpedia

Article excerpt

Introduction

In the last decade important efforts have been made in the development of standards to represent the characteristics (or metadata) of learning objects (LOs) with the objective of facilitating its sharing and reuse among different learning designs and also among different educational tools (Wiley, 2000). Two standards stand out over others, Dublin Core (Hillmann, 2005), oriented to the modeling of any digital resources, and IEEE LOM (IEEE LTSC, 2002), developed specifically to represent the pedagogical features of the resources used in learning activities.

In parallel with this effort, the technology of LO repositories has been developed, which led to the apparition of frameworks, architectures, and protocols for the publication, querying and reuse of LOs through its metadata (Massart, 2006; Ternier et al., 2009). LO repositories can be characterized from several perspectives (McGreal, 2008), among which we highlight (i) whether they contain generic LOs that cover a number of different pedagogical subjects or domains; and (ii) whether they model the semantics of LO metadata through an ontology. These semantic repositories are particularly interesting because they improve the interoperability among educational tools, support semantic queries about the of the repository content, and facilitate the automatic generation or enrichment of some metadata (Soto Carrion et al., 2007). This paper focuses on this last point and specifically on how the categories of the LOs might be automatically generated or annotated through ontologies.

The description of LO metadata through ontologies and the development of techniques for its automatic generation has attracted great interest in recent years (Pahl & Holohan, 2009). In (Jovanovic et al., 2006) authors propose to use two different ontologies for representing LOs: one for describing their content and structure, and the other one for modeling their categories. In this work the annotation of categories is a semi-automatic process that requires the participation of users to manually select the domain concept at which the LO belongs to. In (Dodero et al., 2005) an authoring collaborative environment that allows incorporating domain ontologies to the vocabulary managed by the users in the annotation is presented. In (Al-Khalifa & Davis, 2007) authors have developed an ontology-based architecture to generate LOs from folksonomies that users create manually through tagging. The annotation process consists of mapping the tags of the folksonomy to the instances of the ontologies used to describe semantically the LO domain. In (Nesic et al., 2011) a set of ontologies are defined to represent the domain, social context, and structure of LOs, which are documents shared in a social network. In this work, users manually create the LO metadata through a graphical interface that generates the RDF instances of the ontologies. However, the previously described approaches share the same drawbacks:

* They use ontologies that describe the semantics of a particular domain, making impossible to apply these approaches to the creation/maintenance of cross-domain LO repositories.

* The annotation process is at most semi-automatic, requiring users to annotate and validate the correctness of the metadata. However, it is not feasible to apply this strategy to large-sized repositories.

To address these drawbacks some approaches that generate automatically LO metadata based on cross-domain ontologies that contain a large number of instances (densely populated) have appeared. Niemann & Wolpers (2010) describe semantically the named entities extracted from the manual annotation of LOs (locations and persons) with instances of DBpedia (Bizer et al., 2009a). In Svensson et al. (2010) the contextual information, mainly geographical, associated to LOs created with mobile devices is also annotated with DBpedia. However, authors do not detail whether users manually carry out this annotation or not. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.