Parkinson's disease prediction using voice signals

Parkinson's Disease (PD) is a chronic neurological disorder that gradually impairs movement, speech, and motor coordination as a result of the degeneration of dopamine-producing neurons in the brain.Detecting PD at an early stage is difficult because noticeable motor symptoms usually emerge only after significant neurological damage has occurred.However, changes in speech and voice often appear earlier and can serve as useful indicators of the disease.Recent developments in machine learning and signal processing have enabled the creation of automated and objective diagnostic systems.Voice-based analysis has become a promising, non-invasive technique for Parkinson's detection, allowing meaningful acoustic features to be extracted from speech signals.These features capture subtle variations in vocal production that are not easily observed through clinical examination alone.This project proposes a Parkinson's Disease prediction system using voice signal processing and machine learning models.Acoustic attributes such as pitch, jitter, shimmer, Mel-frequency cepstral coefficients (MFCCs), and harmonicto-noise ratio are derived from voice recordings and used to train classification algorithms.The system aims to deliver an affordable, reliable, and accessible diagnostic support tool for early screening and remote monitoring, thereby assisting clinicians in improving patient care.