基于时间序列相似性分析探讨中风方剂中的症—药演变规律

Exploration of the Evolvement Rule of Symptoms-medicines in the Prescriptions for Stroke with Similarity Analysis of Time Series

【目的】基于中医古籍智能检索平台,采用时间序列相似性分析方法探索治疗中风方剂的症-药演变规律,为中风的临床诊疗提供参考。【方法】筛选治疗中风的方剂,对中风相关症状与治疗中风方剂的药物进行提取与标准化,依据方剂所属古籍的成书或刊刻时间,形成时间序列数据,依据形态相似距离评估不同历史时期的症-症、药-药、症-药间相似性,从而探寻治疗中风方剂中症-药的演变规律。【结果】共收集公元341-1949年间的2065首治疗中风的方剂,涉及中医古籍258本。症-症相关性分析得到3类有关中风的症状群;全时期的症-药时间序列中共记录中药643味,频次居前100位的中药可聚为4类中药群。症-药相关性分析结果显示症状群Ⅰ(真中风症状群)与中药群Ⅰ相关性最显著,症状群Ⅱ(类中风症状群)与中药群Ⅱ相关性最显著,症状群Ⅲ(急中风症状群)与中药群Ⅰ相关性最显著。【结论】通过中风症-药时间序列相似性分析发现,中风相关症状逐渐形成真中风、类中风与急中风3组症状群,与其密切相关的中药可分为4类,各症状群与各中药群关系迥异,其中药配伍日益丰富,治法组合也逐渐完善。挖掘所得结果可为中医药治疗中风的理论发展和临床应用提供一定的参考。

Objective Based on the intelligent retrieval platform of ancient Chinese medical books(http://114.255.40.130:62080/home),similarity analysis of time series was carried out to explore the evolvement rule of symptoms-medicines in the prescriptions for the treatment of stroke, so as to provide reference for clinical treatment of stroke. Methods The prescriptions for the treatment of stroke were screened. The symptoms of stroke and the herbs in the prescriptions for the treatment of stroke were extracted and standardized,and the time series data were formed based on the written time or publication time of the ancient books which recorded the prescriptions. The similarities of symptoms-symptoms, medicines-medicines and symptoms-medicines in different historical periods were assessed based on the morphological similarity distance,and then the evolvement rule of symptoms-medicines in the prescriptions for the treatment of stroke was explored. Results A total of2 065 prescriptions for the treatment of stroke recorded in the year of 341 to 1949 AD were collected,and the prescriptions were distributed in 258 ancient Chinese medical books. The symptoms-symptoms correlation analysis yielded three categories of symptom groups related to stroke;a total of 643 herbal medicines were recorded in the symptoms-medicines time series of the whole period, and the top 100 herbal medicines in terms of frequency could be clustered into four categories of herbal medicine groups. The results of the symptoms-medicines correlation analysis showed that symptom groupⅠ(true stroke symptom group)had most significant correlation with herbal medicine groupⅠ,symptom groupⅡ(apoplectoid stroke symptom group)had most significant correlation with herbal medicine group Ⅱ, and symptom group Ⅲ(acute stroke symptom group) had most significant correlation with herbal medicine groupⅠ. Conclusion The time series similarity analysis of symptoms-medicines for stroke revealed that the symptoms related to stroke gradually develop into three groups of symptoms,i.e.,true stroke symptom group, apoplectoid stroke symptom group and acute stroke symptom group, and the herbal medicines for stroke can be divided into four groups. The relationship between the symptom group and the herbal medicine group varies greatly,with the compatibility of herbal medicines being increasingly rich and the complex therapies gradually improved. The results of the analysis may provide a reference for the theoretical development and clinical application of Chinese medicine for the treatment of stroke.