Building an allergens ontology and maintaining it using machine learning techniques

Populating an Allergens Ontology Using Natural Language Processing and Machine Learning Techniques

“…The work presented in this paper extends our ontology population process, presented in [20], with a 2nd processing stage, that of extracting properties and relations. The problem of properties and relations extraction was examined in the 7th Message Understanding Conference [10] where it was formulated as a separate extraction task.…”

“…Our approach [20] populates the ontology with new instances located in domain specific corpora and also enriches it with the typographic variants of its instances (different lexicalizations of the same instance) located again in the same corpora. The key idea behind this approach is that we can keep the instances of the domain ontology and their typographic variants up-to-date in a semi-automatic way, by periodically re-training an information extraction system using a domain specific corpus.…”

“…The above problems motivated us to build a formally defined logic-based ontology for the allergen domain [20] which could be machine exploitable. We were also concerned with the efficient maintenance of such an ontology.…”

“…As it is presented in [20] we used well established principles and tools to design and build our ontology. We use the description logic-based ontology language, OWL [14] since logic-based languages promise more reusability, hence less cost on ontology building.…”

“…We have used well established tools and design principles to build an ontology in the allergens domain based on previous work and allergen related resources [20]. In this work, we present our approach for the automation of ontology population exploiting the allergens ontology built and PubMed [13] abstracts on the allergens domain and using natural language processing and machine learning techniques.…”