The paper “Linking chemical and disease entities to ontologies by integrating PageRank with extracted relations from literature”, co-authored by Pedro Ruas, André Lamúrias and Francisco M. Couto has been published at Journal of Cheminformatics, a top-ranked journal (h5-index 44).

The paper proposes REEL (“Relation Extraction for Entity Linking”), a framework to improve graph-based named entity linking models with the help of relation extraction tools. For each text document containing chemical or disease mentions, the model generates a disambiguation graph with the potential ontology candidates for the mentions and then extracts relations present in scientific literature to improve the semantic information in the graph. At the end, each mention in text is disambiguated to the ontology concept that best captures its meaning in the respective context.

The model can be integrated in information extraction pipelines, which extract relevant information from large amounts of text, like the case of scientific literature. The potential benefits include better access to scientific information by researchers and the improvement of the curation process of semantic resources (knowledge bases, ontologies, terminologies). The paper is available here.