Multiple Features for Clinical Relation Extraction: a Machine Learning Approach.

•Clinical relation extraction aims to identify semantic relations between entities.•We propose a model with a novel set of knowledge and sentence embedding features.•Experiments on two benchmark datasets from MADE 2018 and n2c2 2018 shared tasks.•The context between entities plays a crucial role in relation detection.•State-of-the-art results are improved by 3.5% on electronic health record notes.