FAIRClinical D3.1: The specification of the meta-data enrichment model

To support patient clinical data discovery and analysis across scientific publications, basic data such as patients demographic data, diseases, and medications need to be identified and harmonized. Scientific publications, however, are available in the form of unstructured text, making much of information not readily available for further reuse and analysis. Therefore, there is an immediate need for automated approaches to extract knowledge from this growing body of literature. To develop and establish reliable machine learning systems, many task specific datasets have been created and made publicly available. Here, we examine the available datasets and their applications for the development of our machine learning approaches for information extraction from clinical case reports.