Clinical Trials for Evidence-Based Medicine in Spanish (CT-EBM-SP) Corpus and word-embeddings

<p>The NLPMedTerm project aims at providing the research community with resources for natural language processing (NLP) in the health domain for Spanish.
A Work Package of the project is a corpus texts annotated with medical entities as a resource for experiments in Named Entity Recognition. The corpus is aimed at training machine-learning models incorporating state-of-the-art neural network approaches.
Another Work Package are the word embeddings from the medical domain.</p>

<p>
The Clinical Trials for Evidence-Based-Medicine in Spanish (CT-EBM-SP) corpus is a collection of 1200 texts about clinical trials studies and clinical trials announcements:
- 500 abstracts from journals published under a Creative Commons license, e.g. available in PubMed or the Scientific Electronic Library Online (SciELO)
- 700 clinical trials announcements published in the European Clinical Trials Register and Repositorio Español de Estudios Clínicos

The word-embeddings were trained with fastText and using the following parameters: skipgram model, window size = 10, dimensions = 100, minimum frequency = 1, number of negatives sampled = 10, learning rate = 1e-4.
We used texts from the European Medicines Agency corpus (∼13.9M tokens) and articles from the Scientific Electronic Library Online (SciELO) repository (∼25M tokens)</p>