Biomedical Named Entity Recognition based on Long and Short Term Memory Model

DOI

10.2991/mcei-17.2017.113How to use a DOI?

Keywords

Long and short term memory model; Literature mining; Biomedical naming recognition; Neural network

Abstract

In view of the problem of biological entity recognition, this paper proposes an improved Long Short-Term Memory (LSTM) recognition method based on the improved bidirectional long and short term memory model. First of all, based on the improvement of re constructed corpus is used to solve the imbalance problem in the distribution of biological entities data sampling algorithm; then, by coupling the forgotten and the input threshold combination to improve the LSTM memory unit, update method to choose the reasonable use of forgotten door control unit state in memory left information improve the biological entity recognition effect. Finally, the test was carried out on the JNLPBA 2004 corpus, and the accuracy rate of 79.7% and the value of 74.1% of the F were obtained. Experimental results show that the proposed recognition method not only has better generalization ability without external assistance, but also effectively improves the recognition effect of biological entities.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Youliang Huang
AU  - Sajid Ali
AU  - Li Wang
AU  - Renquan Liu
PY  - 2017/12
DA  - 2017/12
TI  - Biomedical Named Entity Recognition based on Long and Short Term Memory Model
BT  - Proceedings of the 2017 7th International Conference on Mechatronics, Computer and Education Informationization (MCEI 2017)
PB  - Atlantis Press
SP  - 536
EP  - 540
SN  - 2352-538X
UR  - https://doi.org/10.2991/mcei-17.2017.113
DO  - 10.2991/mcei-17.2017.113
ID  - Huang2017/12
ER  -