《基于隐结构模型的名老中医诊治慢性支气管炎用药规律探讨》由会员分享，可在线阅读，更多相关《基于隐结构模型的名老中医诊治慢性支气管炎用药规律探讨（18页珍藏版）》请在金锄头文库上搜索。
1、基于隐结构模型的名老中医诊治慢性支气管炎用药规律探讨摘要 ?樘教置?老中医诊治慢性支气管炎的用药规律，该文从构建的现代名老中医肺病著作文献数据库中提取 1 274 例慢支病案和临床经验，先用 Lantern 3.1.2（孔明灯）隐结构分析软件构建药物的隐结构模型并进行隐类诠释，再用 SAS 9.1 对药-药、药-症、药-证等不同字段间进行关联规则挖掘。通过挖掘发现 1 274 例慢支病案常用药物甘草、苦杏仁、半夏、陈皮、茯苓，药物类型止咳平喘药、补气药、清化痰热药，药味甘、辛、苦，归肺、脾、胃经，药性温、平、微温，常用有名方有小青龙汤、麻杏石甘汤、二陈汤；构建一诊 147 种药物隐结构模型得到
2、 44个隐变量，88 个隐类，7 个综合聚类模型；药-药二联有 9组、三联 50 组、四联 89 组；对 297 例有两诊的病案经过诊次间的用药比较，得到药-症关联 24 组，二诊减少的药物有麻黄、苦杏仁、半夏等，增加的有茯苓、白术、党参等；药-证二联有 20 组，三联 8 组；对一诊前 83 位症状和前 96 味药物建模得到 50 个隐变量，101 个隐类。名老中医治疗慢支时常用苦杏仁、半夏、陈皮、茯苓，药性偏温、偏甘，刚柔并济，归肺、脾、肾经；常用药组有白术和茯苓，细辛、五味子和半夏，茯苓、党参、半夏和陈皮。药症（证）关联有麻黄、苦杏仁、桂枝和痰少，半夏和外寒内饮证、痰浊阻肺证、肺脾两虚证
3、，苦杏仁和痰浊阻肺证、痰热蕴肺证、风寒犯肺证。隐结构模型可量化中医隐性知识，首次扩展至药物、药症层面，量化出不同变量在同一组数据中所处的主、次、兼地位。该研究得出的名老中医辨治慢支的规律对临床诊疗有一定的参考价值，其数据来源、分析方法、挖掘结果相对可靠、客观，可为其他疾病辨治规律的挖掘提供参考。 关键词 隐结构模型；慢性支气管炎；名老中医；医案；用药规律；数据挖掘 Medication rules of famous veteran traditional Chinese medicine doctor in treatment of chronic bronchitis based on i
4、mplicit structure model CHEN Li-ping1，2，CAI Yong-min3*，LI Jian-sheng3 （1. Zhang Zhongjing College of Chinese Medicine，Nanyang Institute of Technology，Nanyang 473004，China； 2. Institute of Literature in Chinese Medicine，Nanjing University of Chinese Medicine，Nanjing 210046，China； 3. Henan Provincial
5、Synergistic Innovation Center for Respiratory Disease Diagnosis and Treatment and New Drug Research and Development，Henan University of Chinese Medicine，Zhengzhou 450046，China） Abstract To explore the medication rules of famous veteran traditional Chinese medicine （TCM） doctors in treatment of chron
6、ic bronchitis，a structured medical record database for famous veteran TCM doctors in modern clinical books was established. First，Lantern 3.1.2（Kongmin light） implicit structure analysis software was used to build an implicit structure model and make an implicit interpretation. Then，SAS 9.1 was adop
7、ted to mine herb-herb，herb-symptom and herb-syndrome association rules. Through the mining，1 274 commonly used herbs for chronic bronchitis were found，including liquorice，bitter almond，pinellia，dried tangerine or orange peel，poria cocos. The medicine types included antiasthmatic medicine，qi-tonifyin
8、g medicine，and heat-phlegm removing medicine. The medicine tastes included sweet，pungent and bitter. The meridian distributions included lung，spleen and stomach channels. The famous commonly used prescriptions included Xiaoqinglong decoction，Maxing Shigan decoction and Erchen decoction. The 147-herb
9、 implicit structure model for the first diagnosis was built to get 44 hidden variables，88 hidden classes，7 comprehensive clustering models，9 dual herb associations，50 triple herb associations and 89 quadruple associations. Totally 297 medical records for the second diagnosis were compared to obtain
10、24 herb-symptom associations，which reduced ephedra，bitter almond，pinellia and added poria cocos，atractylodes，dangshen，20 dual herb associations and 8 triple herb associations. A model for the top 83 symptoms and top 96 herbs in the first diagnosis was built to get 50 hidden variables and 101 hidden
11、classes. The commonly used herbs of famous veteran TCM doctors included bitter almond，pinellia，dried tangerine or orange peel，poria cocos，which feature mild property and sweet favor and enter lung，spleen，kidney meridians；the commonly herbal pairs included atractylodes and poria cocos，asarum，fructus
12、schisandrae and pinellia，as well as poria cocos，dangshen，pinellia and dried tangerine or orange peel. The herb-syndrome associations involved ephedra，bitter almond，cassia twig and scanty sputum，pinellia and external cold and internal fluid syndrome，turbid phlegm obstructing lung syndrome and spleen
13、deficiency syndrome，bitter almond and turbid phlegm obstructing lung syndrome，phlegm-heat accumulated in the lung and wind-cold invading the lung. The implicit structure model can be used to quantify tacit knowledge of TCM，extend to the herb and syndrome level for the first time，and quantify the maj
14、or，minor and compatible statues of different variables in the same data. This study concluded that the medication rules of famous veteran TCM doctors in treatment of chronic bronchitis have certain reference value for the clinical diagnosis and treatment. The data sources，analysis methods and mining
15、 results were relatively reliable and objective，and can provide a reference for rules of other disease treatment based on syndrome differentiation. Key words implicit structure model；chronic bronchitis；famous veteran traditional Chinese medicine doctor；medical record；medication rules；data mining 为寻找
16、名老中医临床诊治肺系疾病的有关规律，以国家临床研究基地建设项目构建的现代名老中医肺病著作文献数据库中有关慢支病案的数据为研究对象，采用隐结构分析和关联规则的挖掘方法，探讨名老中医对慢支的组方用药规律，为临床辨治该病提供参考。 1 材料与方法 1.1 数据采集 提取肺系著作数据库1内的慢性支气管炎病案和临床经验 1 274 例，含 269 种医籍，385 名医家。先?h 除数据库内与分析无关的变量，如书目 ID，文献 ID等，再将字符型变量按统计软件对数据的要求转变成数值型，并对数据变量进行筛选和规范，以保证统计结果的准确性，如同一证型的不同描述规范为一个证型，如风寒袭肺证、风寒伤肺证、外感风寒证、风寒伤表证风寒犯肺证。一药多名定 1 个名称，如金不换、田七、三七参三七；拆分合写药名，如焦三仙（拆）焦山楂、炒神曲、炒麦芽；因产地或炮制不同而致药物名称不同，规范为原药材名，如潞党参党参、炙紫菀紫菀；同一药物的不同药用部位进行合并，如瓜蒌壳、瓜蒌皮、瓜蒌仁、瓜蒌衣瓜蒌，但因药用部位不同功效发生
