| Objective:Explore the new methods of mining the main syndrome differentiation rules of physicians to treat insomnia,and on this basis,develop a preliminary expert simulation system to initially realize the prescription inference function.Methods:This study will be divided into three parts,the integrated use of complex network knowledge mapping technologies such as data mining method combing physicians treating insomnia the main rule,through computer programming and related development tools,converting the main rule to inference rules,prescription form expert simulation system,finally,prospective study specific cases testing system function is as follows1.A total of 204 subjects(349 consultations)were included in the retrospective study,among which 35 high-frequency symptoms were extracted,and 40 high-frequency drugs were used for data mining by decision tree method to establish a decision tree classification model for doctors to distinguish and treat the main symptom of insomnia,providing a basis for the main reasoning rules of the expert system2.By introducing the visualization analysis method of knowledge map and comprehensively applying the techniques of complex network level analysis co-occurrence analysis and module analysis based on community detection method,the retrospective case analysis was carried out,and the rules and empirical prescriptions for doctors’ differentiation and treatment of insomnia were obtained from different perspectives,providing support for the detailed reasoning rules of expert system.3.On the basis of the above research findings,the simplified expert simulation system is designed and developed by using computer programming tools to realize the preliminary simulation function of square root.51 prospective cases were collected for a small sample test of system function.Jaccard coefficient was used as a similarity evaluation method to test prescription similarity and similarity ratio.The relationship between prescription similarity and curative effect was also explored.Results:This study included a total of 204 cases of retrospective studies,a total of 349 consultations,including 35 kinds of high-frequency symptoms and 40 high-frequency drugs;a total of 51 cases of prospective studies were included.1.The medicine used by the doctor to treat insomnia is mainly heart,liver and meridian,the four qi is mainly cold,and the five flavors are mainly sweet,bitter and spicy.A total of 8 important attributes were selected:sweating,red tongue,fat tongue,binocular dryness,anxiety,greasy fur,bitter mouth,and low mood;the overall decision tree model obtained a total of decoction and Danzhi Xiaoyaosan prescriptions,Xiaochaihu Tang Recipe,Chaiqin Wendan Decoction Recipe 9 categories of judgment rules;the specific decision tree models of various parties obtain a total of 13 specific classification judgment rules.The classification accuracy specificity sensitivity of each model is good,which can provide the main recommendation rules for the construction of expert system.2.The co-occurrence analysis suggests that the physician’s experience in treating insomnia is composed of salvia miltiorrhiza,sour jujube kernels,raw keel,raw oysters,and silk tree skin.Based on the module analysis of community detection method,a total of four symptom modules and four drug modules are obtained,and there is a certain correspondence between the drug and disease modules,which reveals the "pathogenesis-treatment method-prescription"implicit in the drug and disease information.Physicians treat insomnia with depression,heat(fire),and phlegm(wet)as the core pathogenic factors of disturbing the heart.Relieving depression,clearing heat,and reducing phlegm are the main treatments;Zhengxu loses energy to support Qi deficiency,blood deficiency,and Yin deficiency.As the main pathogenesis,replenishing qi,nourishing blood and nourishing yin are the core treatments for deficiency syndrome.If the heat syndrome is significant or the yin deficiency is internal heat,the main agent is Lily Dihuang Decoction,Lily Zhimu Decoction,consistent decoction,Erzhi Wan and other agents.If the symptoms of phlegm dampness or phlegm heat are obvious,the physician tends to add or subdue Wendan Decoction,such as Chaiqin Wendan Decoction or Huanglian Wendan Decoction.Those with less heat and accompanied by symptoms of spleen deficiency,such as partial stool,are more likely to be treated with Danzhi Xiaoyaosan,Xiaoyaosan,etc.;If the signs of heat and phlegm and dampness are not obvious,it is often treated by reconciliation and law,and the prescriptions are mostly added and subtracted by Bupleurum preparations,such as Xiao Chaihu soup and Chaihu Longgu oyster soup.Physicians attach importance to the adjustment of the Qi machine,and apply the Qi medicine more frequently.3.An expert simulation system has been preliminarily constructed,based on logical rule reasoning to realize the function of simulating prescription.In the system function test,there are 28 prescriptions with a similarity coefficient greater than 0.3,accounting for 54.9%,which meets the accuracy standard set by this study(50%).It can be considered that the simulated prescription of this system is basically similar to the actual prescription of doctors,and can preliminarily reflect the characteristics of the law of doctors’ treatment based on syndrome differentiation.Conclusion:1.The combination of data mining,knowledge graph,and complex network technology can discover the main laws of doctors’ syndrome differentiation and treatment,and can be applied as an effective method for mining syndrome differentiation experience;2.The expert simulation system formed based on the integrated method has certain accuracy and potential application value,and is worthy of further development and research;3.The doctor’s diagnosis and treatment rules and the characteristics of prescription medication can provide reference and experience for clinical diagnosis and treatment of insomnia. |
PDF Full Text Request