Research And Realization Of TCM Auxiliary Diagnosis And Treatment System For Chronic Obstructive Pulmonary Disease

Chronic Complications of Pulmonary Disease(COPD)is a common and frequent chronic respiratory disease with high morbidity and mortality,and great harm to human health.Its prevention has become a major public health problem.Traditional Chinese medicine emphasizes holistic view and differentiation of symptoms and signs,and has the advantages of clear curative effect and small side effects in the management of chronic diseases of COPD.The base of COPD patients in China is large,but the resources of TCM diagnosis and treatment are relatively insufficient.Therefore,the auxiliary diagnosis and treatment of TCM for COPD has important theoretical and practical significance.First,aiming at the problem of syndrome identification in the auxiliary diagnosis and treatment of COPD,a syndrome classification algorithm based on a deep belief network and a syndrome classification algorithm based on a deep feedforward network are proposed in this thesis,the two algorithms are used to determine the syndrome category according to the patient’s symptoms.The syndrome classification algorithm based on deep belief network uses the powerful feature extraction function of deep belief network,and extracts effective high-dimensional information from symptom features through pre-training of unsupervised learning.The syndrome classification algorithm based on the deep feedforward network calculates the Fisher score to eliminate invalid features,and then uses the deep feedforward network to complete the classification task.Secondly,in response to the problem of drug recommendation in the auxiliary diagnosis and treatment of COPD,a method of COPD TCM recommendation is proposed in this thesis based on the correlation between symptoms,syndrome types,and drugs.The task of drug recommendation is divided into two parts: typical drug recommendation and supplementary drug recommendation.The typical drug recommendation part uses frequency analysis and PSO-BP neural network to construct a typical drug recommendation model for each syndrome type,and supplementary drug recommendation is implemented based on the association rules mined by the Apriori algorithm.After the typical drugs and supplementary drugs are tested for the compatibility of traditional Chinese medicines,a reliable drug recommendation result can be obtained.Finally,a COPD-oriented TCM auxiliary diagnosis and treatment system is designed and implemented in this thesis.The system has the functions of standardized information collection,auxiliary syndrome differentiation,intelligent drug recommendation,and medical record data management,which can assist doctors in the process of COPD patients’ treatment,so as to achieve the purpose of improving the efficiency of diagnosis and treatment.Experiments’ results show that the syndrome classification and drug recommendation methods proposed in this thesis exhibit better performance than other methods on the COPD TCM diagnosis and treatment data set of the Department of Pulmonary Diseases,Provincial Hospital of Traditional Chinese Medicine.The accuracy rate of syndrome classification is 82.39%,and the recommended mean square errors of the nine syndrome types of typical drugs are all less than 0.01,indicating that the method of syndrome classification and drug recommendation in this thesis have a good auxiliary diagnosis and treatment effect.