A Survey of Machine Learning Approaches for Parkinson’s Disease Prediction

Parkinson’s disease, a degenerative disorder of the nervous system, has a significant impact on a substantial population across the globe, leading to gradual deterioration in both, non-motor and motor, capabilities. Timely identification and intervention are of utmost importance in managing the symptoms and enhancing the quality of patients’ lifestyle. This paper presents a comprehensive exploration of Parkinson’s disease, focusing on its symptoms, diagnosis, machine-learning applications, and predictive models. It delves into various machine learning techniques applied to predict and identify Parkinson’s disease, leveraging datasets and algorithms to discern patterns and relationships for early detection and treatment. The study extensively reviews existing literature, highlighting key research papers, methodologies, and findings pertaining to the utilization of machine learning in Parkinson’s disease diagnosis and prediction. It covers a range of techniques employed in different studies, such as Random Forest, K-Nearest Neighbor, Convolutional Neural Networks, Support Vector Machine, and others. These papers utilize diverse datasets and algorithms, examining features like vocal characteristics, gait analysis, brain morphological changes, hand movement signals, Electroencephalography signals, and more to diagnose or predict Parkinson’s disease. The aim emphasizes the promising potential of machine learning in aiding early detection and personalized treatment plans for Parkinson’s disease.