http://jcps.bjmu.edu.cn

中国药学(英文版) ›› 2024, Vol. 33 ›› Issue (2): 110-122.DOI: 10.5246/jcps.2024.02.010

• 【研究论文】 • 上一篇    下一篇

基于数据挖掘分析中药治疗头风病的用药规律

张啸, 吴春兴, 王博龙*()   

  1. 宜春学院, 江西 宜春 336000
  • 收稿日期:2023-10-15 修回日期:2023-11-26 接受日期:2024-01-08 出版日期:2024-03-03 发布日期:2024-03-03
  • 通讯作者: 王博龙

Exploring dosage patterns of Chinese herbal medicine for recurrent headaches: A data mining analysis

Xiao Zhang, Chunxing Wu, Bolong Wang*()   

  1. Yichun University, Yichun 336000, Jiangxi, China
  • Received:2023-10-15 Revised:2023-11-26 Accepted:2024-01-08 Online:2024-03-03 Published:2024-03-03
  • Contact: Bolong Wang

摘要:

本研究基于数据挖掘的方法, 探究中药治疗头风的用药规律, 为该病的临床治疗提供参考。首先收集相关数据库及经典中医书籍中主治疾病为头风的中药方剂, 然后运用频次分析、关联规则分析、聚类分析、决策树分析等方法探究中药治疗头风的组方规律。经过筛选共得到有关方剂174首, 涉及中药227味, 累计使用1622次。结果显示治疗头风的中药多属解表类, 活血化瘀类和补虚类, 以性温味辛入肝经或性温味辛入脾经为主; 核心中药主要有川芎、白芷、细辛、防风、甘草等。关联规则分析共得到中药组合92项; 聚类分析得到6个治疗头风病的聚类方; 决策树分析得到白芷为最佳识别中药, 且CHAID算法的准确率最高。本研究通过数据挖掘发现了头风治疗的核心中药, 揭示了方剂中潜在的组方规律, 明确头风治疗的遣方用药模式, 为该病的临床诊治提供了参考。

关键词: 头风, 数据挖掘, 用药规律, 聚类分析, 决策树模型

Abstract:

This study was carried out based on data mining methods aimed at exploring the medication patterns of Chinese medicine in the treatment of recurrent headaches, thereby providing valuable references for the clinical treatment of this condition. Initially, relevant databases and classical traditional Chinese medicine (TCM) books containing prescriptions for treating recurrent headaches were gathered. Subsequently, various techniques, such as frequency analysis, association rule analysis, cluster analysis, and decision tree analysis, were employed to investigate the prescription patterns of Chinese medicine for recurrent headaches. Following the screening process, a total of 174 prescriptions were obtained, encompassing 227 Chinese medicinal ingredients that were utilized cumulatively 1622 times. The findings indicated that Chinese medicine commonly employed in treating recurrent headaches predominantly fell under the categories of dispersing wind, promoting blood circulation and removing blood stasis, and tonifying deficiency. These medicines often possessed warm and pungent flavors and had an affinity for the liver meridian or spleen meridian. Key Chinese medicine identified included Chuanxiong rhizoma, Angelicae dahuricae radix, Asariradix et rhizoma, Saposhnikoviae radix, and Glycyrrhizae radix et rhizoma. Furthermore, through association rule analysis, a total of 92 combinations of Chinese medicine were identified. Cluster analysis yielded six distinct clusters of prescriptions for treating recurrent headaches. Simultaneously, decision tree analysis demonstrated that Angelicae dahuricae radix emerged as the most influential drug, with the CHAID algorithm exhibiting the highest accuracy rate. This study successfully employed data mining techniques to identify the core Chinese medicine used in treating recurrent headaches. It unveiled potential prescription patterns within these prescriptions and clarified the medication strategies for managing this condition. Consequently, these findings provided valuable insights for clinical diagnosis and treatment of recurrent headaches.

Key words: Recurrent headaches, Data mining, Drug administration rules, Cluster analysis, Decision tree model

Supporting: