Predictive Modeling for Parkinson’s disease Diagnosis: A Machine Learning Approach

Parkinson's disease is a progressive brain degeneration disorder that highly influences the lives of patients, hence detection at early stages becomes the utmost necessity. The current research article develops the ML pipeline for the classification of prediction with a patient dataset using vocal characteristic. We have implemented SVM algorithm, Random Forest, and KNN algorithms, and optimized model performance applying hyperparameter optimization and k-fold cross-validation. Results reported that the optimal SVM model yielded the maximum accuracy, indicating the possibility of an earlier diagnosis in clinical practice. This study focuses on the utilization of ML to make the invasive methods of diagnosis more effective for Parkinson's disease prediction and thus provides directions for future work.