Volume - 6 | Issue - 4 | december 2024
Published
02 January, 2025
Parkinson's Disease (PD) is a progressive neurodegenerative disorder that impacts motor skills, including tremors, bradykinesia, and rigidity, affecting millions globally. Early diagnosis is essential for effective treatment yet remains challenging as the symptoms overlap with other conditions and the limitations of conventional diagnostic methods. This study presents a diagnostic tool utilizing machine learning that employs a Support Vector Machine (SVM) classifier for precise PD prediction through biomedical voice data. The system uses the UCI Parkinson’s dataset, where pre-processing tasks like feature standardization, train-test split (80-20 ratio), and Recursive Feature Elimination (RFE)enhance model accuracy by identifying significant features. An easy-to-use Streamlit web application was developed to enable real-time predictions, permitting users to input voice parameters and receive instant diagnostic results. The SVM classifier achieved a precision rate of 92%, showcasing its capability and effectiveness in distinguishing PD from non-affected cases. By providing a scalable, cost-effective, and non-invasive approach, this tool bridges advanced computational techniques with real-world healthcare needs. Future enhancements will focus on integrating multimodal data, such as neuroimaging and wearable sensor data, as well as employing deep learning models to improve diagnostic accuracy and expand clinical applicability.
KeywordsParkinson’s Disease Machine Learning Voice Biomarkers SVM Classifier Diagnostic Tool Streamlit Application
