Application Of Optimized BP Neural Network In The Classification And Diagnosis Of Acute Chest Pain

Chest pain is a common emergency in the emergency department.The acute chest pain disease is urgent,the treatment time window is narrow,and the differential diagnosis is difficult,which brings great challenges to the diagnosis of emergency department doctors.Early evaluation of acute chest pain patients with neural network can assist doctors in diagnosis and optimization of the treatment process,to win valuable treatment time for patients.In this paper,a classification diagnosis system based on five typical acute chest pain diseases was designed to meet the needs of doctors for rapid diagnosis of acute chest pain diseases.The relevant characteristics of each of the five chest pain diseases were analyzed and extracted to form a sample data set,of which 120 groups were used for modeling and 20 groups were used for testing.The sample data were fitted by the basic BP neural network and the BP neural network model optimized by genetic algorithm,and the preliminary diagnosis results were obtained by the calculation of several parameters of the patient’s clinical characteristics.The specific research work is as follows:(1)BP neural network for diagnosis of acute chest pain was constructed.The influencing factors of five kinds of acute chest pain diseases were analyzed,and sample data were extracted and integrated to construct a basic BP neural network.The network trained by the sample data could predict the type of disease suffered by the new samples.(2)Genetic algorithm is used to optimize the initial weight threshold of BP neural network.The random weight threshold of BP neural network makes the prediction result unstable.The optimized BP neural network can improve this problem,improve the accuracy of prediction,and make the BP neural network applied in the diagnosis of acute chest pain more reasonable.(3)A visual user interface was designed to construct a diagnosis system for acute chest pain.The system operation process is simple,can display the preliminary diagnosis result accurately,has the practicability,is easy to popularize.The design of this system relies on MATLAB software,the construction of the system model is realized by MATLAB programming language,genetic algorithm toolbox,etc.,the visual user interface of the system is designed by GUI,a graphical user interface tool.The test results of 20 new samples show that the prediction accuracy of BP neural network optimized by BP neural network and genetic algorithm is 85% and 95% respectively,indicating that the system is effective and has the use value.