Diagnosis of Parkinson Disease Patients Using Egyptian Vulture Optimization Algorithm

Abstract

Parkinson disease(PD) is a neurological disorder which affect the nervous system of the body causing problem related to gait and speech disorder. Speech and gait serve as major parameter in diagnosis of the disease in the early stages of its symptoms. This study uses these parameters to perform a comparative study of two nature inspired algorithm for diagnosis of the Parkinson Disease. The process involves first selecting an optimal feature set for classification and then using them to classify and predict PD patients from non PD patients. Two different datasets were used consisting of gait and speech data of PD and non PD patients. Optimal Feature selection was done using Particle swarm optimization and Egyptian Vulture Optimization Algorithm. The optimal feature set was then used to classify the dataset using KNN classifier. According to the experiment EVOA outperforms PSO in the selection of the feature subset. This study thus concludes that new meta-heuristic algorithm EVOA works better than traditional PSO in diagnosis of PD patients which in real life can help to speed up the process and lessen the suffering of the patient by early detection.