Evaluation of linked, open data sources for mining adverse drug reaction signals

Abstract

Linked Data is an emerging paradigm of publishing data in the Internet, accompanied with semantic annotations in a machine understandable fashion. The Internet provides vast data, useful in identifying Public Health trends, e.g. concerning the use of drugs, or the spread of diseases. Current practice of exploiting such data includes their combination from different sources, in order to reinforce their exploitation potential, based on unstructured data management practices and the Linked Data paradigm. In this paper, we present the design, the challenges and an evaluation of a Linked Data model to be used in the context of a platform exploiting social media and bibliographic data sources (namely, Twitter and PubMed), focusing on the application of Adverse Drug Reaction (ADR) signal identification. More specifically, we present the challenges of exploiting Bio2RDF as a Linked Open Data source in this respect, focusing on collecting, updating and normalizing data with the ultimate goal of identifying ADR signals, and evaluate the presented model against three reference evaluation datasets.