摘要

This paper proposes a staged and distributed strategy for biomedical named entity recognition (Bio-NER). In this strategy, named entities (NEs) are recognized in two stages: dictionary-based NE tagging and machine learning-based NE recognition. In dictionary-based tagging stage, we exploit a distributed strategy to recognize NEs, in which different models and NE dictionaries are used for recognizing different NEs. This strategy not only exploits strengths of dictionary-based model and machine learning-based model, but also improves the performance and speed of identification. Experimental results show that our approach achieves higher accuracy than previous approaches.