Advancing Parkinson's Disease Detection: An Empirical Evaluation of Machine Learning Models Based on Speech Analysis

Parkinson's disease (PD) is a neurodegenerative disorder that adversely impacts a considerable portion of the population, especially senior citizens. Currently, the basis of PD diagnosis is clinical assessments, which can be costly, time-consuming, and invasive, that can be subjective or prone to errors. Dysarthria, a prevalent condition characterized by slow and distorted speech, often coexists with PD, presenting an opportunity to leverage speech features for diagnostic purposes. In this research paper, the altered speech features caused by Dysarthria are harnessed to train a diverse set of cutting-edge machine learning algorithms, including Artificial Neural Networks (ANN), Support Vector Machines (SVM), Random Forests, and Decision Trees. A thorough examination of the latest advancements in the diagnosis of Parkinson's disease (PD) using speech analysis is provided. Additionally, a comparative analysis between the performance of a newly introduced classifier, HyperTab, and existing models is conducted. The findings reveal the remarkable promise of machine learning in speech-based PD diagnosis, paving the way for developing a cost-effective tool for accelerating disease detection. Moreover, this contribution is crucial in providing vital support to those areas with limited access to specialized medical facilities, significantly improving the quality of life and outcomes for PD patients in rural regions.