Voice signal analysis for early detection of Parkinson's disease using machine learning techniques

The early detection of Parkinson's disease (PD) is crucial for its effective management and treatment, as it can significantly slow disease progression and improve quality of life. One promising approach for early diagnosis is the analysis of voice signals, which can reveal subtle changes in phonetic features associated with PD. This study explores the use of machine learning techniques to identify PD at an early stage by leveraging a dataset from the UCI Machine Learning Repository, consisting of 147 phonetic samples from PD patients and 48 from healthy controls. The methodology involved preprocessing the data, selecting relevant features using a genetic algorithm, and addressing class imbalance with the Synthetic Minority Oversampling Technique (SMOTE). Principal Component Analysis (PCA) was employed for dimensionality reduction, followed by the application of Support Vector Machines (SVMs) and k-Nearest Neighbor (KNN) classifiers. Cross-validation was performed to evaluate model performance. The results indicate that the KNN classifier achieved the best accuracy of 96.11%, demonstrating its superior capability in distinguishing between PD patients and healthy individuals based on voice features. The high accuracy suggests that voice signal analysis, combined with advanced machine learning techniques, is a promising avenue for the early detection of Parkinson's disease. This research underscores the potential of non-invasive diagnostic tools in clinical settings, paving the way for further studies to refine and validate these methods for broader applications.