Voice-Driven Diagnostics: Leveraging Machine Learning for Early Detection of Parkinson's Disease from Acoustic Biomarkers

Parkinson's disease (PD) is globally the fastestgrowing neurodegenerative disease in the elderly with unclear causes that impairs individuals' motor usage and may induce dementia. Using machine learning technology has the potential to allow for faster and more accurate diagnoses of Parkinson's, which can allow Parkinson's treatment to reach low-income communities with access to fewer health resources. Our study explores a machine learning and deep learning classification algorithm that uses voice recordings to diagnose whether or not a person is likely to have Parkinson's disease. By using various characteristics of the voice recording, such as pitch, frequency, and amplitude, our model is able to accurately predict the presence of Parkinson's with an accuracy of 94%. Unlike other contemporary methods of diagnosing Parkinson's, our research is more accessible as recording voices can be done universally and is not dependent on access to healthcare technology.