Research on the Construction and Optimization of TCM Word Network for TCM-clinical Decision Support System
10.19879/j.cnki.1005-5304.202311427
- VernacularTitle:面向中医临床辅助决策系统的中医药词网构建及优化研究
- Author:
Shanrong HUANG
1
;
Ling ZHU
1
;
Shaolei TIAN
1
;
Wanting ZHENG
1
;
Yuanyuan TONG
1
;
Jinghua LI
1
Author Information
1. 中国中医科学院中医药信息研究所,北京 100700
- Publication Type:Journal Article
- Keywords:
TCM word network;
TCM-clinical decision support system;
construction method
- From:
Chinese Journal of Information on Traditional Chinese Medicine
2025;32(2):58-63
- CountryChina
- Language:Chinese
-
Abstract:
Objective To discuss the method of constructing TCM word network and the subsequent optimization method;To solve the problems of natural language processing in the TCM-clinical decision support system using TCM terminology.Methods This study was based on several national standards and terminology dictionaries of TCM,such as Zhong Yi Lin Chuang Zhen Liao Shu Yu Ji Bing Part and Zhong Yi Yao Xue Zhu Ti Ci Biao,supplemented with related corpus on the Internet,and the framework of TCM word network was constructed with ontology as the guide.Word2Vec was used to load pre-trained models for similarity calculation,statistical and rule-based new word discovery,bibliometric analysis,and other methods to add or delete concept words and synonyms such as symptom terms and prescription terms.Results The TCM word network contained 55 semantic types and 13 semantic relationships,more than 110 000 concept words,and more than 300 000 synonyms.In addition to classifying concept words by semantic types,more detailed semantic classification was carried out in the upper and lower tree structures,in which more than 30 000 concept words had superordinate words,and nearly 2 600 concept words had subordinate words.At the same time,a process-based terminology supplementation mechanism,update review mechanism and cleaning mechanism were created.Conclusion TCM word network can not only be applied to the natural language processing process of TCM-clinical decision support system,but also can be applied to the normalization of data and the automatic construction of knowledge graph.
