Recognising Named Entity of Medical Imaging Procedures in Clinical Notes
Ruan Wei, Won‐Sook Lee · 2018
Information on medical imaging procedures in free-text clinical notes plays a significant role in help diagnosis process. In this paper, we present a named entity of the medical imaging procedure recognition system based on conditional random fields (CRF) model with word-based, part-of-speech, Metamap semantic and et.al features. The system is trained and tested on a manually labelled dataset from I2B2 with the F1score up to 0.923 for recognizing medical imaging procedure entities. Our system can be customized by simply defining new medical named entities, which has been proved on medication recognition with the F1score from 0.870 to 0.937.