Detection of Parkinson’s disease using Neural Network Trained with Genetic Algorithm

Abstract

Recent research works have focused on detection of Parkinson’s disease using several machine learning techniques. Accurate
separation of normal persons in the subjects under consideration from the persons being affected by Parkinson’s disease is a challenging job. In
the present work Neural Network (NN) has been trained using Genetic algorithm (GA employed to detect persons being affected with
Parkinson’s disease. The initial weight vector to the input layer of the NN has been optimized gradually using the optimization techniques to
enhance the performance of NN to a greater extent. The experimental results of the proposed method have been compared with a well-known
Multilayer Perceptron Feed-Forward Network (MLP-FFN) and also with the NN. Performance measures like accuracy, precision, recall and Fmeasure
have been used to compare the performances of the algorithms. The experimental results have revealed significant improvement over
the existing performances to detect Parkinson’s disease using GA.

Keywords: Parkinson’s Disease; Genetic Algorithm; ANN; MLP-FFN; Gradient Descent