HITSZ _ CDR System for Disease and Chemical Named Entity Recognition and Relation Extraction

In this paper, an end-to-end machine learning-based system was proposed for the challenge task of chemical and disease named entity recognition (DNER) and chemical-induced diseases (CID) relation extraction in BioCreative V, where DNER includes chemical and disease mention recognition (CDMR) and normalization (CDN). The system consists of six components: a preprocessing module, two individual sequence labeling module, an ensemble module, a normalization module and a CDR extraction module. The two sequence labeling modules and the ensemble module were designed for CDMR. Evaluation using the challenge corpus showed that our system achieved the highest F1-scores of 86.76% on CDMR, 67.82% on CDN and 41.26% on CID relation extraction, respectively.