Named Entity Recognition Based on Transformer Encoder in the Medical Field
Xiyv Mou, Xindong Zhang · 2022 3rd International Conference on Electronic Communication and Artificial Intelligence (IWECAI) · 2022
With the large-scale popularization of electronic medical record system (electronic medical record system EMR), EMR contains lots of consultation, examination and other information about patient. So how to make full use of these data to provide support for follow-up medical research and medical practice has developed as a research focus in the current medical field. Named entity recognition in medical texts is the basic task among them. The existing medical named entity recognition models are mostly founded on the BiLSTM neural network. Because the BiLSTM network is difficult to parallel calculations, and there is a risk of gradient disappearance or explosion, this paper designs a Transformer Encoder medical named entity recognition model combined with BERT. The mechanism for extracting features enables the network to parallel calculations and at the same time improve the recognition effect.