A heuristic approach with artificial neural network for Parkinson’s disease

Abstract

Parkinson’s disease is a common neurological disorder. Its symptoms are more commonly in the form of motor misfunctioning. As the disease progresses, non-motor symptoms are also observed. In previous studies, feature selection methods have been used and shown significant benefits in the diagnosis of the Parkinson’s Disease in patients. Feature selection methods aim to improve the classification performance by eliminating non-valuable or less-valuable features. In this study, we aim to analyze, for diagnosing Parkinson’s disease, the voice recordings of the patients with applying a recent bio-inspired optimization technique namely the Wolf Search Algorithm (WSA). WSA is a bio-inspired heuristic optimization algorithm which has been inspired by the natural behavior of wolves in daily life. We also use an artificial neural network model with feature selection methods, for the purpose of classification of the Parkinson’s Disease in patients. We investigate the classification performances of the combinations of WSA-based feature selection method with well-known feature selection methods namely Information Gain and ReliefF feature selection methods. Experimental results show that ReliefF feature selection method outperform than the other feature selection method combinations for the diagnosis of the Parkinson’s Disease in patients.