Research On The Mining Method Of Prescription Rules Based On Medical Record Data

Traditional Chinese Medicine(TCM)has received increasing attention for its great potential in preventing and treating diseases and improving health,which has brought about a growing demand for Chinese medicine data resources.During the development of Traditional Chinese Medicine,a diagnosis and treatment system has been formed.This system is based on the theory of yin and yang and the theory of the five elements.It uses inspection,auscultation,olfaction,inquiry,and palpation to identify the patient’s symptoms,identify diseases,and choose appropriate treatment programs.For thousands of years,TCM doctors have accumulated many medical records with proven curative effects.Therefore,the medical case data preserved in ancient literature and clinical applications are valuable resources for tacit mining knowledge and guiding the clinical practice of TCM.At present,data mining methods for TCM cases mainly include frequency statistics method,association rule method,cluster analysis method,factor analysis method,artificial neural network method,and so on.These methods are only for the order of drugs.They do not study the relationship between medications and symptoms and do not fully dig out hidden laws and potential knowledge.Therefore,it is crucial to introduce new technologies and methods to quickly extract commonly used drugs and drug combinations from many medical case data.This combination has a definite curative effect in treating diseases.At the same time,determine the core drugs that play a crucial role in the disease process,and analyze the relationship between specific symptoms and medicines.It is hoped that researchers can make full use of the current experience prescriptions,promote the dissemination and development of medical workers’ academic thinking,and promote and apply their clinical experience.ObjectiveA new method for mining prescription patterns of TCM medical case data is studied based on TCM medical case data characteristics,combined with network topology mining algorithms.Medical cases of TCM clinical treatment of headache are collected and collated as research examples.Their prescription patterns are analyzed through three steps:standardization of medical case information,construction of complex networks,and mining of prescription patterns to illustrate the specific application of the new method.MethodsThis research takes the data mining literature of Chinese medicine cases as the starting point,systematically reviews the application of commonly used data mining techniques in the research scope of Chinese medicine records,points out the shortcomings,and puts forward a prospect.A new method of mining prescription rules based on medical record data is designed and established to improve some of the current problems.At the same time,a set of headache medical case data was excavated and analyzed.The results were cross-validated using relevant literature and expert knowledge bases related to the fundamental theories of traditional Chinese medicine and modern pharmacology to illustrate the application of the new method in drug compatibility.Results1.This study designs and establishes a method for mining prescription medication patterns based on TCM medical case data.This study makes full use of empirical prescriptions.It innovatively introduces the PageRank algorithm for deeper mining of drug-symptom networks,which will eventually lead to a new method of drug ranking.We can summarize the effective drugs,common drug combinations,and symptom-drug relationships in the prescription case-cohort with the new plan.This approach will help uncover hidden patterns and provide a methodological example for the transmission of clinical experience.2.Clinical cases of headache treatment in Chinese medicine were used as research examples to verify the feasibility of the new method.Rule 1:The 21 driver herbs are Chuanxiong,Gancao,Baishao,Fuling,Danshen,Danggui,Gouteng,Tianma,Shengdi,Juhua,Chaihu,Gegen,Baizhi,Shengjiang,Muli,Chishao,Chenpi,Zexie,Dazao,Baizhu,and Manjingzi.Rule 2:The standard drug combinations are Danggui-Chuanxiong,Gouteng-Tianma,Baizhi-Chuanxiong,Baizhu-Danggui,Gouteng-Baishao,Chishao-Danshen,Fuling-Baizhu,Tianma-Duzhong,Quanxie-Danshen,Zaoren-Tianma.Rule 3:Symptom-drug combinations are Eye distension-Gouteng,Tight pulse-Shengjiang,Rough pulse-Wugong,Dreaminess-Zaoren,Palpitation-Gegen,Flusteredness-Danshen,lumbar soreness-Shengdi,Dreaminess-Tianma,Fullness in head-Gouteng,Insomnia-Shengdi.Besides,statistical analysis has been added to understand the distribution characteristics and nature of symptoms and drugs in headache medical records.The medicinal properties of Chinese herbs are mainly cold,followed by warm and flat medicines,with a few being cool and hot.Herbs’ taste is primarily bitter and sweet,followed by savory,salty,sour,light,and finally astringent.In terms of meridians,these herbs enter the liver and lung meridians the most.Conclusion1.This study uses network topology mining algorithms to explore the prescription pattern mining method more comprehensively based on medical case data from three perspectives.This method has strong practicability,can pass on experience,and can also provide specific methodological examples for medical case mining.2.The medical records of Chinese medicine clinical treatment of headaches are taken as research examples,and the prescription rules contained in the excavation reveal the primary information in three aspects:Firstly,Chuanxiong is the first choice in the treatment of headache;secondly,the core medicines are primarily found in combination,and the treatment methods embodied are consistent with the primary treatment methods of the conventional medicine of headache diseases,which are to dispel wind and free the network vessels,calm the liver and extinguish wind,quicken the blood and transform stasis,and supplement vacuity;thirdly,specific medicines are chosen according to particular symptoms to adjust the taste and dosage of the treatment.The third is to select specific drugs based on specific symptoms and adjust the dosage in time.The new method of prescription pattern mining based on medical case data has been verified to have more reliable clinical practice significance and popularized application value.