Early Detection of Parkinson's Disease as a Pre-diagnosis Tool Using Various Classification Techniques on Vocal Features

Parkinson's disease is a non-curable progressive nervous system disorder affecting operations related to muscles movements. More than 1 million individuals are affected by Parkinson's disease in India per year. Due to the progressive nature of the disease, the symptoms generally start gradually and are barely noticeable; with symptoms that can start from normally unnoticeable shaking of a hand to noticeable speech and writing changes to even worse like loss of automatic movements. This project is concerned with contributing to the advancement of medical technologies and may help earlier detection of Parkinson's which will enable early treatment. In this paper, we have overviewed the current status of Parkinson's disease detection and studied the model for early detection of Parkinson's disease using various classifier approaches. The highest accuracy of about 96.61% was achieved using the XgBoost classifier.