Computer-aided Diagnosis Of SLE And Analysis Of Significant Influencing Factors

Systemic Lupus Erythematosus(SLE)is one of the most common autoimmune diseases in rheumatology and immunology.Lupus nephropathy is the most common complication of SLE and can cause severe internal organ failure.The timely detection,diagnosis and control of SLE with renal involvement can play a positive role in the treatment of SLE patients,which is conducive to the implementation of treatment decisions,improve efficacy,and reduce mortality.Two jobs were completed in this article:Classification-assisted diagnosis of SLE with renal involvement was resolved through the use of a Simulated Annealing-Support Vector Machines(SA-SVM)algorithm model.The improved Relief-F algorithm was proposed to analyze the relevant influencing factors in patients with confirmed SLE with renal involvement.First,the computer-aided diagnosis system has not been universally adopted for SLErelated conditions.This article builds a SA-SVM classification model for SLE with renal involvement.Simulated Annealing(SA)algorithm optimization supports The penalty coefficient C and the kernel function parameter g in the Support Vector Machines(SVM)achieve the purpose of parameter optimization and classification-assisted diagnosis in the SVM.Then,for the issues such as insufficiency of the factors affecting SLE with renal involvement,an improved Relief-F algorithm is proposes in this paper to evaluate and screen the influencing factors of this condition.Finally,SA-SVM classification model is used to verify the effectiveness of the improved algorithm.The experimental results show that SA optimization for SVM improves the accuracy of a single SVM classifier.It is able to predict the number of cases in batches,and the highest classification accuracy rate is 98.72%.It is proved that the classification model in this paper has a good potential in the diagnosis of SLE diseases,and can provide necessary methods and experimental references for the study of computer-aided diagnosis system for this disease.Based on the improved Relief-F SLE with renal involvement and non-concurrent renal disease data,the characteristic weights were evaluated and their significant influencing factor selection results were validated in both the corresponding medical findings and in the SA-SVM classification experiments.It shows that improving the Relief-F is more helpful to obtain the factors that affect the significant differences between the two types of diseases,and it has certain advantages over the pre-improvement algorithm.In summary,I hope that the research in this paper can provide a certain methodological reference for the construction of computer-assisted diagnosis system for SLE-related diseases and the analysis of significant factors;Helps discover,diagnose and treat SLE and its complications in clinical medicine.