基于关联规则的中医药知识发现——以肺病医案为例  

Knowledge Discovery Based on Association Rules——Taking Pulmonary Disease Medical Cases as an Example

作　　者：马捷[1,2] 王馨梓 何美慧 Ma Jie;Wang Xinzi;He Meihui(School of Business and Management,Jilin University,Changchun 130022,China;Information Resource Research Center,Jilin University,Changchun 130022,China)

机构地区：[1]吉林大学商学与管理学院,吉林长春130022 [2]吉林大学信息资源研究中心,吉林长春130022

出　　处：《现代情报》2024年第3期34-46,共13页Journal of Modern Information

基　　金：吉林大学“中国式现代化道路”与“人类文明新形态”哲学社会科学研究创新团队项目“推进人类数字生态文明:数字化转型视域下的数据价值与数据创新”(项目编号:2022CXTD20)。

摘　　要：[目的/意义]中医医案作为重要的医疗资源含有大量的数据与信息,深度挖掘中医医案,围绕医案内容进行知识发现,有利于隐性知识的发掘,提高中医临床诊疗的效率与医疗服务的精准性。[方法/过程]以实地肺病医案为数据源,基于统计分析与关联规则的方法对中医医案内容进行挖掘,以咳嗽为主证为出发点,挖掘相关中医用药规律与临床症状以及症—药规律,探究关联规则分析在中医药知识发现方面的创新应用。[结果/结论]本地治疗咳嗽的高频药物包括炙甘草、杏仁、半夏、川贝母、茯苓、陈皮等,舌诊症状以苔薄白、苔薄黄、舌红、舌暗红、有齿印等为常见。基于Apriori算法的关联规则分析显示常见的药物配伍为丹参,桔梗—炙甘草、柴胡,桔梗—炙甘草、前胡,枳壳—桔梗等;舌诊症状以薄淡黄—舌暗红、有齿印—舌暗红、苔薄腻淡黄—有齿印等组合为强关联组合,苔薄淡黄—舌暗红症状组合的强关联药物有桔梗、炙甘草、丹参、白前等。咳嗽患者的类型有外感咳嗽与内伤咳嗽,涉及风热证、痰湿证、阴虚证、肺气虚证、痰热壅肺证等证型。[Purpose/Significance]As an important medical resource,TCM(Traditional Chinese Medicine)medical records contain a large amount of data and information.Deep mining of TCM medical records and knowledge discovery around the contents of medical records are conductive to the mining of tacit knowledge,improving the efficiency of TCM clinical diagnosis and treatment and the accuracy of medical services.[Method/Process]The study took the medical records of local lung disease medical records as the data source,the content of traditional Chinese medicine medical records was mined based on statistical analysis and association rules.Starting from cough as the main symptoms,the relevant traditional Chinese medicine medical patterns,clinical symptoms,and symptom drug patterns were explored,attempted to explore the innovative application of association rule analysis in discovering traditional Chinese medicine knowledge discovery.[Result/Conclusion]The local high-frequency drugs for cough include Radix Glycyrrhizae Preparata,almond,Pinellia ternata,Fritillaria cirrhosa,Poria cocos,dried tangerine peel,etc.The common symptoms of tongue diagnosis are thin white coating,thin yellow coating,red tongue,dark red tongue,tooth printed tongue,etc.Association rule analysis based on the Apriori algorithm shows that common drug combinations include Salvia miltiorrhiza,Platycodon grandiflorum-Radix Glycyrrhizae Preparata,Bupleurum chinense,Platycodon grandiflorum-Radix Glycyrrhizae Preparata,peucedanum root,Fructus aurantii-Platycodon grandiflorum.The symptoms of tongue diagnosis are dark red tongue with thin yellow coating,dark red tongue with tooth marks,thin greasy yellow fur coating with tooth marks and other combinations as the strongly correlated combination.The strongly correlated drugs of thin yellow fur dark red tongue symptom combination are Platycodon grandiflorum,Radix Glycyrrhizae Preparata,Salvia miltiorrhiza,and Cynanchum glaucescens.The types of cough patients include the type of external cough and internal injury cough,which

关 键 词：关联规则 知识发现 APRIORI算法 中医医案 用药规律 

分 类 号：G250.7[文化科学—图书馆学]