Detecting Parkinson's Diseases with MVDP and Ml Classification Algorithms

Abstract

The motor system is the main target of the neurodegenerative disease known as Parkinson's Disease (PD). Degeneration of brain neurones caused by intricate interplay between genetic and environmental factors, rather than infectious or contagious dissemination, is the cause of this disease's spread. Researchers have begun to use various machine learning techniques to detect and assess Parkinson's disease (PD) utilising audio input and MRI/PET or DAT scans due to the increase in the number of individuals with the disease. In order to identify Parkinson's disease, we need to create a system that can analyse patients' auditory data and identify patterns in their conduct. In order to construct the disease-detecting classifier, the suggested method makes use of Support Vector Machine, random Forest, and a number of other algorithms. A preprocessing step and data analysis are utilised to manage data, guarantee an adequate level of detection error, and optimise training time. This data is later combined with other datasets for use in training and testing. With soft voting, our model achieves a final accuracy of 94.87% and an F1 score percentage of 96.9%.