The Application Of Artificial Intelligence In The Diagnosis Of Diabetes

Diabetes has become one of the major diseases affecting the health of our country's citizens,and the incidence is increasing.At present,the number of diabetic patients and undiagnosed patients in China are still the highest in the world.However,problems such as uneven distribution of medical resources,shortage of doctors,and lack of experience of young doctors have severely constrained the diagnosis of diabetes.Therefore,it is very necessary to help doctors diagnose diabetes by means of science and technology.The development of artificial intelligence technology is chan ging with each passing day.Alpha Go's victory in man-machine war and the powerful cognitive computing capabilities of Watson robots have aroused people's high attention and imagination for artificial intelligence.Artificial neural network is the forefront interdisciplinary subject with rapid development of artificial intelligence in the world.It has strong learning and computing power,and can better adapt to the change of data space.Its application opens up a new way for the research of artificial intel ligence.Based on this,this thesis researched the application of artificial intelligence technology in the diagnosis of diabetes.Establish diagnosis model of diabetes based on BP neural network.The experimental data are extracted and integrated,and inp ut feature vectors are selected to build neural network models.In order to improve learning efficiency,the parameters of the model were adjusted.The feasibility of applying BP neural network to diabetes diagnosis is verified through experiments.Establish diagnosis model of diabetes based on probabilistic neural network(PNN).In order to improve the accuracy and speed of diagnosis,the PNN with simple network structure and concise training is selected to establish diabetes diagnosis model.After comparison of the two models,it is found that PNN needs less parameters to adjust,and does not need to determine the network structure of the hidden layer and the number of hidden layer neurons,which is easier to be implemented and used than the BP network.This thesis calculated the average accuracy and the corresponding standard deviation of the 20 tests,it shows that PNN model is more suitable for building diabetes diagnosis model.Based on the establishment of the diagnosis model,this thesis combined the LabVIEW and MATLAB to design an auxiliary diagnosis system for diabetes.It calls the neural network diagnosis module,completes the diagnosis of diabetes by analyzing the user data,while giving the corresponding diagnosis and treatment advice.Compared with the traditional diagnosis process,it can effectively save the doctor's time and improve the diagnosis efficiency.