From PubMed to ontologies

by Mohammad Halawani

16:00 (40 min) in USB 5.008

We are investigating our work towards building an ontology of rehabilitation treatments using data-driven methods to support the ontology development process. Here, we are discussing retrieving full text literature from PubMed and extractive representative domain terminology, which is used as a data source to bootstrap the ontology.

The literature defines the scope of the ontology. We start from an initially small corpus of literature, expanding this to a set that covers a full domain. Using information extraction, we define an initial terminology that provides the basis and the competencies for constructing the ontology. Out approach advances the ontology development with innovations from the big data analytics.