BioC and Simplified Use of the PMC Open Access Dataset for Biomedical Text Mining

High quality easily-accessible resources are crucial for developing reliable applications in the health and biomedical domain. At the same time, interoperability, broad use, and reuse are vital considerations when developing useful systems. As a response, BioC has recently been put forward as a convenient XML format to share text documents and annotations, and as an accompanying input/output library to promote interoperability of data and tools. The BioC approach allows a large number of different textual annotations to be represented, and permits developers to more easily and efficiently share training data, supportive software modules and produced results. Here we give a brief overview of BioC resources. We also present the BioC-PMC dataset as a new resource, which contains all the articles available from the PubMed Central Open Access collection conveniently packaged in the BioC format. We show how this valuable resource can be easily used for text-mining tasks. Code and data are available for download a