A Study to Evaluate Various Machine Learning Approaches for Early Prediction of Neurodegenerative Diseases

Alzheimer's and Parkinson's diseases are prevalent neurodegenerative disorders among the elderly, often with a genetic predisposition, posing significant challenges in early diagnosis and intervention. Timely detection is crucial to slowing disease progression and improving patient outcomes. This study systematically evaluates the efficacy of various machine learning algorithms for early prediction of neurodegenerative diseases using two distinct datasets: handwriting samples for Alzheimer's and biomedical voice measurements for Parkinson's. Ten widely used machine learning models, including logistic regression, support vector machines, random forest, decision trees, K-nearest neighbor, autoencoder, hybrid model, Gaussian mixture model, naïve Bayes, and gradient boosting, were as-sessed for their predictive performance. The results indicate that the random forest model consistently outperformed others for both Alzheimer's and Parkinson's by achieving 91% and 94% accuracy respectively. These findings highlight the potential of machine learning techniques in enhancing early diagnostic tools for neurodegenerative diseases.