Machine Learning Frameworks for Effective Diagnosis of Parkinson’s Disease Using NB, GNN, and GBM Algorithms

Parkinson's Disease (PD) is a neuro degenerative disorder affecting millions of patients globally, causing motor impairments like tremors & stiffness, impacting on daily activities and quality of life.Parkinson's disease arises when dopamine producing neurons in the substantia nigra, a region of the midbrain, disrupting the normal functioning of the basal ganglia.This neuronal loss leads to difficulties in speech, writing, walking and performing everyday tasks.As the condition progresses, symptoms worsen and nonmotor issues such as cognitive decline, mood disorders and sleep disturbances often emerge.The frameworks investigates the potential of Machine Learning (ML) algorithms in predicting PD.Machine learning algorithms like Naive Bayes (NB) Classifier, Graph Neural Network (GNN) and Gradient Boosting Machine (GBM) Protocols are applied to patient's data like demographics, clinical evaluations and potential biomarkers etc. Naive Bayes classifier is a simple but effective probabilistic model that performs well with categorical data and assumes feature independence.Graph Neural Network is a flexible algorithm capable of modeling complex nonlinear relationships in data.Gradient Boosting is powerful ensemble method that iteratively improve predictions by combining weak learners, optimizing for accuracy and minimizing errors.Simulation results shows the performance of the proposed ML algorithms, which significantly enhances prediction of PD in terms of accuracy upto 96.4%, sensitivity of 97.1%, selectivity ranging upto 94.3%, positive and negative predictive values, and F1-score etc.