Research On The Recommendation Of Decocting Time Of Traditional Chinese Medicine Compound Based On The Similarity Of Prescription

Chinese medicine decoction is the most common preparation,the quality of decoction directly affects the clinical therapeutic effect,the correct decocting method can maximize its effect.Ancient Medical Classics recorded in detail the amount of water added and the amount of decocting,and the decocting process was relatively elaborate.In the process of decocting modern TCM decoction,the research of decocting duration and other parameters is still not deep enough.Most decoctions adopt relatively single decocting mode,without considering the characteristics of the prescription and adopting the corresponding decocting scheme.When decoction takes a long time to fry,it is often not fully cooked;When fast frying is needed,it is often overdone,and the effects are often not optimal.With the rapid development of decocting equipment,the method of semi-automatic control of decocting process is out of date.How to use artificial intelligence technology to excavate the ancient famous scientific decocting rules and recommend scientific decocting parameters efficiently for modern intelligent decocting equipment is an urgent scientific problem that needs to be solved.The common parameters of TCM decocting include the amount of water added,the amount of decocting,the time of strong fire and the time of slow fire,etc,among which the duration of simmering is the core of the decocting parameters The recommendation of its parameters solves the key problems of the intelligent decocting equipment of TCM compound and has very important application value and practical significanceWith the wide application of natural language processing technology in the field of traditional Chinese medicine,text similarity of traditional Chinese medicine prescriptions has been deeply studied,and text similarity based method can be used to calculate the similarity of prescription texts.In this study,the ancient famous prescriptions with the greatest similarity to TCM compounds were matched through the similarity calculation of prescriptions,and the decocting duration parameter was recommended as the reference of TCM compounds decocting.In this study,similarity was studied from two aspects of prescription text and multidimensional characteristics:(1)The optimal matching algorithm based on probability model(BM25)was used to match the most similar ancient famous prescription from the prescription text similarity;(2)The formula similarity algorithm in line with the formula composition,dose and functional dimension characteristics,including Jaccard similarity coefficient,cosine similarity algorithm and LDA topic model,was adopted,and multi-dimensional recommended decocting duration parameters were integrated.The experiment showed that the fusion of multidimensional prescription similarity recommendation fully considered the characteristics of the prescription,compared with BM25 algorithm and single dimension recommendation,has a better effectThe main work of this paper is as follows:(1)Build the decocting information model and its database.Firstly,the name,source,composition,dosage,function,decocting method and other information of prescription in classical medical books such as Treatise on Febrile Diseases and Synopsis of the Golden Chamber were analyzed,and the information was divided into multiple levels and the attribute set of each level was established Secondly,data preprocessing,the traditional Chinese medicine prescription text data for structured processing and ancient and modern units of measurement and measurement conversion,using the mainstream conversion rules,the ancient weight,volume and other units of measurement unified into modern units of measurement.Thirdly,by establishing the entity and the correlation model between entities,the cooking information model is constructed Finally,using database technology and structured data after processing,the ancient famous decoction database is constructed(2)Research on decocting time recommendation of TCM compound based on optimal matching algorithm.The Jieba word segmentation tool is used to analyze and construct morpheme,then pre-process the document and query,adjust the parameters,calculate the BM25 score,and return the result according to the score ranking.The weighted average processing of the top five frying times is performed,and the frying time information is the recommended result(3)Research on the recommended decocting time of TCM compounds based on the similarity of fusion multidimensional prescriptions.Following the principle of "similar composition,similar dose and similar function",Jaccard algorithm,cosine similarity algorithm and LDA topic model algorithm were respectively used to match the ancient famous prescriptions with the most similar components,the most similar dose and the same function from the database,and their decocting time information was weighted and fused to recommend the decocting parameters of target TCM compounds.(4)Establishment of decoction recommendation system for TCM compound.Based on Python programming language and Django technical framework,a TCM compound decocting recommendation system was established by using development tools such as Pycharm and MySQL,combined with multi-dimensional formula similarity algorithm,to realize the functions of TCM compound decocting information management and TCM compound decocting recommendation,etc.,providing a management information platform for TCM compound decocting recommendation.