Parkinson’s Disease Detection System Using Machine Learning Techniques

Parkinson’s disease is a chronic neurological disorder that has a major impact on an individual’s motor functioning and has greatly impacted the quality of life due to debilitating symptoms, yet complex etiologies involving genetic and environmental elements. In this work, a speech dataset’s ability to predict Parkinson’s disease using machine learning(ML) approaches is examined. We evaluated eight machine learning(ML) models, namely: Linear Regression (LR), Logistic Regression (LOR), Random Forest (RF), SVM, Decision Tree (DT), XGBoost (XGB), Neural Networks (NN), and AdaBoost (ADB). The challenge of class imbalance has been addressed with the Self-Organizing Map Undersampling (SOMUS) technique. The models were evaluated using Recall, Precision, and F1-score. Our results indicate that the XGBoost and Random-Forest classifiers performed best for parkinson’s disease detection. The study indicates that machine learning can be applied for the improvement of diagnosis at an early stage and intervention in Parkinson’s disease.