Machine Learning-Based Framework for Multi-Class Diagnosis of Neurodegenerative Diseases: A Study on Parkinson's Disease

A new era of intelligent medical diagnostics is emerging with the development of machine learning-based algorithms to diagnose neurodegenerative diseases (NDDs). In the present work, we discuss an innovative framework that uses principal component analysis (PCA) for feature extraction, Fisher discriminant ratio (FDR) for feature selection and support vector machines (SVM) for classification of Healthy controls, Parkinson's Disease and SWEDD subjects. We have extended our framework to handle the challenge of multi-class disease diagnosis, wherein, accuracy up to 100% has been achieved. This demonstrates the potential of the present methodology to be developed into a clinical relevant diagnostic and decision support system.