A Parkinson’s Disease Machine Learning Model for Severity Prediction

Abstract Parkinson’s disease (PD), a neurodegenerative disease, is caused by the degeneration of dopaminergic neurons in the brain. This degeneration affects the normal motor functions of the body. The early diagnosis of the disease is difficult because its symptoms are similar to those of other diseases. However, the severity level of the disease is important for its proper treatment. In this paper, a machine learning framework is proposed for the detection of Parkinson’s disease and the severity level of the disease using a Dopamine Transporter Single Photon Emission Computed Tomography (DAT-SPECT) imaging technique. The framework is based on the Support Vector Machine (SVM) classification technique with a Radial Basis Function (RBF) kernel. The proposed framework is tested on a dataset of DAT-SPECT images and gives an accuracy of 85.5%, precision of 85.8%, recall of 85.2%, and F1-score of 85.5%. These results show the better classification capability of the proposed framework in distinguishing between healthy and diseased people at different severity levels of the disease. Keywords:Parkinson's Disease DAT-SPECT Imaging, Support Vector Machine, ROI Extraction, Severity Prediction, Machine Learning.