Semantic annotation and harvesting of federated scholarly data using ontologies

Effective synthesis of learning material is a multidimensional problem, which often relies on handpicking approaches and human expertise. Sources of educational content exist in a variety of forms, each offering proprietary metadata information and search facilities. This paper aims to show that it is possible to harvest scholarly resources from various repositories of open educational resources (OERs) in a federated manner. In addition, their subject can be automatically annotated using ontology inference and standard thematic terminologies.

Based on a semantic interpretation of their metadata, authors can align external collections and maintain them in a shared knowledge pool known as the Learning Object Ontology Repository (LOOR). The author leverages the LOOR and show that it is possible to search through various educational repositories’ metadata and amalgamate their semantics into a common learning object (LO) ontology. The author then proceeds with automatic subject classification of LOs using keyword expansion and referencing standard taxonomic vocabularies for thematic classification, expressed in SKOS.

The approach for automatic subject classification simply takes advantage of the implicit information in the searching and selection process and combines them with expert knowledge in the domain of reference (SKOS thesauri). This is shown to improve recall by a considerable factor, while precision remains unaffected.

To the best of the author’s knowledge, the idea of subject classification of LOs through the reuse of search query terms combined with SKOS-based matching and expansion has not been investigated before in a federated scholarly setting.