Ontological semantic search of learning objects: experiments and results

Resumo

Generating relevant results in semantic searches typically involves the context of the search term, and may include, besides elements of the search context, other elements like the location, intention, variation of words and synonyms. Ontologies and metadata are important computational tools to bring the semantics for learning object queries. The semantic search system presented in this paper aims to retrieve learning objects organized by OBAA metadata. Its search engine supports the integration of multiple educational ontologies with the OBAA metadata ontology. However, when distinct ontologies need to interoperate, semantic compatibility problems can lead to unpredictable, ambiguous or incomplete results. The semantic search system used in this paper addresses these issues through a combination of ontology aligning and mapping mechanisms. This paper presents the mapping and alignment algorithms used in the search engine. Experiments were conducted to evaluate the quality of the information returned by the search from the point of view of users. These experiments showed positive results.