ATRMiner : A system for automatic biomedical named entities recognition

The recognition of biomedical entities in natural text is a key step for automatic analysis of textual resources. Biomedical entities recognition tools are a prerequisite for many applications working on text. In this paper, we develop a biomedical entities recognition tool named ATRMiner. Using natural language processing technology, ATRMiner can identify five classes biomedical entities, namely gene, disease, cellular components, molecular functions and biological process based on gene ontology. Our system achieved a recall 83%, a precision 92%, and an F-score 87.2% aiming at test datasets. This shows our system is promising for practical application.