The Research And Application Of Improved GA-SVM Algorithm In The Diagnosis Of Secondary Nephritis

Systemic Lupus Erythematosus(SLE)is an unexplained autoimmune disease,which is generally accompanied by a variety of autoantibodies and is co-formed with multiple systemic impairments.Approximately 80%of patients with diagnosed SLE suffer from lupus nephritis(LN).LN is the highest incidence of SLE secondary disease,and because of its complex and diverse causes,many factors have become very difficult to diagnose.Therefore,the diagnosis of LN disease has a very important medical significance.There are two diagnostic methods of medicine,one is diagnosed by diagnostic criteria,one is through the performance of the points for the diagnosis of auxiliary diagnosis,but these two methods there are some drawbacks,the former method prone to misdiagnosis and missed diagnosis,after A method is too cumbersome to use,too much work.Therefore,this paper uses a support vector machine(SVM)model based intelligent algorithm diagnosis method.The method of collecting a large number of LN disease-related data the different cases were expressed in different numbers,through the SVM model for intelligent prediction,and ultimately achieve the purpose of auxiliary diagnosis.However,due to the SVM algorithm in the practical application process,such as the choice of penalty parameters and the selection of nuclear parameters there are some shortcomings.Therefore,this paper presents a model(GA-SVM model)based on Genetic Algorithm(GA)to optimize SVM parameters and kernel function parameters and a GA-SVM optimization algorithm based on LN Based on this,an improved GA-SVM optimization model is proposed.Using three kinds of intelligent algorithm simulation projections for the same data set,respectively,compared the simulation results of three kinds of model,based on the improved GA-SVM algorithm in the error value,mean square error(mse)and linear correlation have made great improvement,more suitable for LN disease diagnosis prediction.