计算机科学与探索 ›› 2017, Vol. 11 ›› Issue (7): 1159-1165.DOI: 10.3778/j.issn.1673-9418.1605036

• 人工智能与模式识别 • 上一篇    下一篇

面向复杂网络的中药方剂配伍规律挖掘算法

韩  楠1，乔少杰2+，李天瑞3，宫兴伟3，舒红平4，元昌安5   

  1. 1. 成都信息工程大学 管理学院，成都 610103
    2. 成都信息工程大学 网络空间安全学院，成都 610225
    3. 西南交通大学 信息科学与技术学院，成都 610031
    4. 成都信息工程大学 软件工程学院，成都 610225
    5. 广西师范学院 科学计算与智能信息处理广西高校重点实验室，南宁 530023
  • 出版日期:2017-07-01 发布日期:2017-07-07

Prescription Compatibility Mining Algorithm of Traditional Chinese Medicine over Complex Networks

HAN Nan1, QIAO Shaojie2+, LI Tianrui3, GONG Xingwei3, SHU Hongping4, YUAN Chang'an5   

  1. 1. School of Management, Chengdu University of Information Technology, Chengdu 610103, China
    2. School of CyberSecurity, Chengdu University of Information Technology, Chengdu 610225, China
    3. School of Information Science and Technology, Southwest Jiaotong University, Chengdu 610031, China
    4. School of Software Engineering, Chengdu University of Information Technology, Chengdu 610225, China
    5. Science Computing and Intelligent Information Processing of Guangxi Higher Education Key Laboratory, Guangxi Teachers Education University, Nanning 530023, China
  • Online:2017-07-01 Published:2017-07-07

PDF

摘要： 针对传统方剂配伍规律分析方法的不足，提出一种面向复杂网络的新型中药（traditional Chinese medicine，TCM）方剂配伍规律挖掘算法。根据中药方剂特性并结合点式互信息构建TCM网络模型，结合TCM网络的小世界特性提出TCM网络的局部适应度模型，分析TCM网络的特性并挖掘TCM网络中配伍关系紧密、相似度较大的药物群。以4 000余首经典方剂作为实验对象，验证了所提方法具有较好的有效性，与经典LFM（local fitness measure）算法对比，平均模块度值提高了0.05，为中药方剂的配伍规律进行探索及新药研发提供了新思路。

关键词: 中药, 数据挖掘, 配伍, 复杂网络, 药物群

Abstract: Aiming to overcome the drawbacks of traditional Chinese medicine (TCM) prescription analysis, this paper proposes a new complex networks-based TCM prescription mining algorithm, which creates the TCM networks by combining the characteristics of prescriptions and point mutual information. This paper also proposes a new local fitness model of TCM networks by integrating the feature of small world, which can analyze the characteristics of TCM networks and discover the closely linked and similar herb groups. Extensive experiments are conducted on more than 4000 prescriptions to evaluate the effectiveness of the proposed algorithm. Compared with the LFM (local fitness measure) algorithm, the results show that the average modularity can be improved by 0.05. The proposed algorithm can be applied to explore the compatibility of prescriptions and provide new ideas for the research and   development of new medicines.

Key words: traditional Chinese medicine, data mining, compatibility, complex networks, herb group

京ICP备13024263号-1
Copyright © 《计算机科学与探索》编辑部
技术支持：北京玛格泰克科技发展有限公司
总访问量
今日访问
在线人数