Machine Learning Approach For Predicting Parkinson’s Disease At Early Stages

Abstract

Parkinson’s Disease (PD) is a chronic neurodegenerative disorder that affects motor control, coordination, and speech. Early detection is essential to slow disease progression and improve life quality. One of the earliest and most prevalent symptoms is dysphonia—altered voice characteristics in terms of pitch, loudness, and quality. In this paper, we propose a predictive framework using the K-Nearest Neighbors (KNN) algorithm for early diagnosis of PD through speech signal analysis. The methodology applies feature extraction, normalization, and dimensionality reduction on vocal parameters such as jitter, shimmer, and harmonic-to-noise ratio. The optimized model identifies PD patients effectively by comparing vocal features with those of healthy controls. The experimental results show that KNN, when properly tuned, offers high accuracy and interpretability for clinical applications.