Developing System-based Voice Features for Detecting Parkinson’s Disease Using Machine Learning Algorithms

Abstract

Parkinson’s disease (PD) is a neurodegenerative disorder characterized by a deficit of dopamine in the brain. This condition has the potential to impact individuals of advanced age. The procedure for diagnosing PD is currently not well established. Diagnostics includes a range of methods, including the identification and evaluation of symptoms, the implementation of clinical trials, and the use of laboratory tests. This research work employs a range of machine learning (ML) algorithms, including k-nearest neighbors (k-NN), support vector machines (SVMs), random forest (RF), logistic regression (LR), and AdaBoost boosting approaches, to predict the occurrence of PD and assist healthcare practitioners in recommending tailored treatment plans. To evaluate the suggested ML methods, it is customary to use a standard dataset consisting of various biological voice measures obtained from individuals afflicted with PD as well as healthy individuals. The experimental results demonstrate that the LR model achieves an accuracy of 86%, the k-NN model achieves an accuracy of 92%, the SVM model achieves an accuracy of 95%, the RF model achieves an accuracy of 95%, and the AdaBoost boosting model achieves an accuracy of 93%. SVM and RF are well acknowledged for their high accuracy in classification tasks. Upon conducting a comparative analysis with other studies, it was shown that the proposed intervention yielded outcomes that were either comparable to or superior to those reported in previous research.