Classification Of Patients With Chronic Spontaneous Urticaria Based On Muti-scale Fuzzy Entroy

Chronic spontaneous urticaria is a common skin disease in clinic,which seriously reduces the quality of life of patients.However,the pathogenesis of urticaria is still unclear.In recent years,many studies have shown that the skin and brain are connected and interact in a two-way manner,it is necessary to excavate the information of brain area for the auxiliary diagnosis of urticaria.Entropy analysis is a useful tool for measuring signal complexity.Using entropy analysis method to analyze functional magnetic resonance imaging with high resolution can more accurately grasp the changes in brain activity in patients with chronic urticaria.This paper analyzes the complexity of brain regions based on entropy as follows:Firstly,the performance of different multi-scale entropy algorithms for feature extraction in brain regions is discussed.Compare the differences in algorithms and performance between MSE and MFE.An analysis is given for the undefined entropy and the selection of MFE membership functions in the MSE calculation process.Apply the two methods to the data in this article and use AUC as the evaluation index to draw a conclusion.For the data in this article,the MFE algorithm is better.Secondly,the MFE algorithm is improved.Although MFE algorithm has great advantages over MSE algorithm,it also has some problems,such as obvious fluctuation of entropy value when the scale is too large and easy to be disturbed by noise.Based on this,this paper proposes an improved multi-scale fuzzy entropy algorithm.In the process of coarsening,the traditional coarsening method is changed into the refined complex coarsening method,which not only solves the problem of inaccurate entropy estimation but also reduces the probability of undefined entropy;For the membership function,the concept of "sensitivity threshold" is introduced and the adjustment factor is added.To a certain extent,it serves the purpose of antinoise interference,improves the stability of the algorithm and is more in line with the physical nature.At the same time,the suitable parameter range of the short data is determined according to the simulation numerical experiment.Finally,the improved multi-scale fuzzy entropy algorithm was applied to the experimental data of this project as a feature extraction method,and the relevant significantly different brain regions were accurately and comprehensively found,which proved the superiority of the improved algorithm.Fisher scores were also used as a feature selection method for f MRI data and were used for classification.It achieves effective differentiation between patients with chronic urticaria and healthy subjects.The brain regions found in this paper are compared with the existing open research,which verifies the effectiveness of the proposed algorithm.