Named entity recognition of diabetes based on ALBERT and BILSTM(PDF)
《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]
- Issue:
- 2021年第11期
- Page:
- 1438-1443
- Research Field:
- 其他（激光医学等）
- Publishing date:
Info
- Title:
- Named entity recognition of diabetes based on ALBERT and BILSTM
- Author(s):
- MA Shiyu; HUANG Runcai
- School of Electrical and Electronic Engineering, Shanghai University of Engineering and Technology, Shanghai 201620, China
- Keywords:
- Keywords: diabetes named entity recognition a lite BERT bidirectional long short-term memory network conditional random field
- PACS:
- TP183；TP391；R319
- DOI:
- DOI:10.3969/j.issn.1005-202X.2021.11.021
- Abstract:
- Named entity recognition of diabetes is able to identify useful key information from the literatures related to diabetes, which is helpful for the diagnosis and treatment of diabetes. A named entity recognition method based on a lite BERT (ALBERT) and bidirectional long short-term memory network is proposed for solving the problems of BERT such as single semantics and limited vocabulary. In addition, the shortcomings of time consuming and high resource cost of dynamic word vector training are also improved. The experiment is carried out on the diabetes data set and then compared with the existing mainstream models. The results show that the entity recognition effect of the model with ALBERT is higher than that of the existing mainstream models, and that the training speed of ALBERT is faster than that of BERT.
Last Update: 2021-11-27