Investigation of machine learning and deep learning methods on voice data in diagnosis of Parkinson's disease

Today, especially with the developments in machine learning, there has been an increase in the studies carried out within the scope of voice disorders and speech measurements. Especially for diseases with speech impairment symptoms such as Parkinson's, classification studies are carried out by using different methods and analysis of sounds. When the studies for Parkinson's disease are examined, it is seen that the studies are generally focused on the evaluation of sounds and patient handwriting with machine learning methods. The aim of this study was to evaluate the performance of machine learning methods using the "Parkinson's Disease Classification (PDC)" dataset created by Şakar et al. (2019). The difference of the study from other studies is to test the success of these techniques on the PDC dataset by using more machine learning and deep learning techniques. According to the algorithm results for the PD classification problem in the literature, it is seen that higher accuracy rates are reported in this study. As a result of the study, it is seen that the best results are obtained from the Convolutional Neural Networks (KSA) algorithm. The KSA technique, which has shown promising results in the PH classification problem in future studies, can be used to estimate the Unified Parkinson's Disease Rating Scale score of PH patients to create a PH tele-monitoring system.