Research And Implementation Of Disease Diagnosis Method Based On Fuzzy Theory And Machine Learning

In recent decades,although medical diagnosis has made great progress,it is difficult for even experienced professional doctors to predict the emergency diseases accurately.With the development of machine learning and deep learning technology,the relevant intelligent algorithms have begun to help doctors identify and predict diseases.However,these data-based methods only train models based on historical data to obtain the required results.They are not based on professional medical knowledge,and they can not get high prediction accuracy in the actual medical datasets.Generally,the relationship between patient and symptoms,the relationship between symptom and disease and even the related medical data are often vague,inaccurate,and uncertain.Therefore,even if the patient information has been clearly provided,how to evaluate specific symptoms or diseases accurately is challenging,which need to be dealt with by trained and experienced doctors clinically.In order to solve the above difficulties in the medical assistant diagnosis system,it is necessary to develop a more intelligent diagnosis model by using fuzzy theory and machine learning methods to overcome the disadvantages of using a certain kind of method alone.This thesis will revolve around the combination of fuzzy theory and machine learning algorithms to solve medical diagnosis problems.The main contributions of the thesis are as follows:(1)The performance of medical diagnosis based on the fuzzy set,intuitionistic fuzzy set,and the neutrosophic set is studied,and a medical diagnosis method based on the combination of neutrosophic recommendation algorithm and spectral clustering algorithm is proposed,which improves the performance of disease diagnosis.(2)This thesis proposes a disease diagnosis method based on feature extraction of the fuzzy neural network model,which is used for doctors to make an initial judgment according to patients' symptoms and effectively helps inexperienced doctors to make diagnoses.The model combines the advantages of fuzzy logic and neural network to make the network have the ability to deal with fuzzy information directly.The dimension of fuzzy features is reduced by the feature extraction method,which improves the accuracy of disease diagnosis and reduces the amount of calculation.(3)A small medical assistant diagnosis system is designed and implemented.The system mainly includes medical data management,model training,and disease diagnosis process.The availability of the system is verified by experiments on four medical datasets.At the same time,doctors can choose different model training data and compare the diagnosis results.