Research On Aided Diagnosis And Management Strategy Of Diabetes Based On Deep Learning

In recent years,the number of patients with chronic diseases in China ranks first in the world,diabetes and its complications as an important part of which seriously affect people's daily lives and pose a serious threat to human health.In the traditional medical diagnosis,the accumulated clinical experiences of doctors in combination with laboratory or instrumental indicators are mainly considered as a diagnostic basis,which easily causes misdiagnosis or missed diagnosis,delaying the timing of treatment.In order to solve the shortcomings of the traditional methods effectively,we propose to use deep learning techniques to build diabetes-related prediction models and analyze them in depth to assist medical staff in decision-making and make diabetes-related management strategies perfect.Through the literature research,diagnosis and treatment features and data of diabetes and its complications are summarized and collected,and deep learning-related model and algorithms are summed up and compared.Based on the two diabetes related data of Pima Indians Diabetes and Diabetic Retinopathy Debrecen,three kinds of techniques,BP neural network,support vector machine and deep belief network,are used to construct diabetes and its complications prediction models for the disease predictive research in order to assist the medical service personnel in making diagnosis and treatment decisions.Then,the classification accuracy,sensitivity and specificity are selected as evaluation criteria to analyze the performance of the three prediction models.The experimental results show that the application of deep belief network in the two diseases is superior to the other two algorithms.At last,the weight matrix of each layer of the optimal disease prediction model based on deep belief network is used to calculate the relative strength values between each input attribute and the output target of the model,and the influence of every factor in the two diseases is systematically analyzed.Through the above analysis,for Type II diabetes,three kinds of management strategies,namely,diabetes prevention education,personal prevention and community prevention,are proposed.For diabetic retinopathy,health education,multi-department alliance and regular eye examination management strategies are proposed to improve the management and control status of diabetes and its complications.