Research On The Quantitative Classification Method Of Gait Features In Patients With Parkinson’s Disease Based On EEMD

Parkinson’s disease is a common neurodegenerative disease with a low mortality rate but a high disability rate,causing endless suffering to patients and family members.At present,both domestic and overseas researchers have conducted a lot of research on the detection method of Parkinson’s disease,and have obtained a series of research results.Among them,the judgment of the severity of Parkinson’s disease is a prerequisite for treatment.According to the different conditions,there are differences in the dosage of drugs and treatment methods.Therefore,how to classify Parkinson’s disease efficiently and accurately is of great significance.However,in the clinical diagnosis of Parkinson’s disease,doctors usually use the scale to classify Parkinson’s disease,which is subjective and inaccurate.This paper quantifies and classifies the gait features of patients with Parkinson’s disease based on EEMD algorithm,constructed a quantified model of gait features of patients with Parkinson’s disease,and verified the model based on relevant experiments.The main contents of this paper are as follows:Decomposition of gait signals in patients with Parkinson’s disease based on EEMD.Due to the unstable symptoms of Parkinson’s disease,this paper takes the COP track as the research object.The T-test determines that the COP track offset has a significant difference.Based on the EEMD algorithm,COP track offset is decomposed to obtain the IMF component with real physical meaning.Gait features extraction and selection based on EEMD.This paper extracts the characteristic parameters of the above effective IMF components and energy entropy,and select the gait feature set through saliency analysis.Parkinson’s disease classification often uses traditional gait feature parameters.Through the correlation experiment between the target level and individual differences,it is verified that the gait feature set based on EEMD has a certain correlation with the target level and is not easily affected by individual differences.Determination of quantitative classification results of gait features in patients with Parkinson’s disease.This paper selects the best number of gait features based on the mRMR and SVM algorithm,and then constructs a gait quantification model for patients with Parkinson’s disease.This paper compares the accuracy of BP neural network,SVM algorithm and PSO-SVM for Parkinson’s disease grading experiments.The result verifies the effectiveness and feasibility of the quantitative classification method of Parkinson’s gait features.In clinical medicine,it can assist the diagnosis of the severity of Parkinson’s disease with the scale.Focusing on the quantitative classification of gait features in patients with Parkinson’s disease,the research on the gait feature extraction,selection and classification of patients with Parkinson’s disease has been carried out.The difficulties are as follows: First,gait signals were determined by analysis of typical motor symptoms of Parkinson’s disease and independent T test methods;Secondly,considering the characteristics of gait signals comprehensively,it is decomposed based on the EEMD algorithm and the gait features are extracted.