A Parallel Machine Learning Framework for Detecting Alzheimer’s Disease

This paper proposes a parallel machine learning framework for detecting Alzheimer’s disease through T1-weighted MRI scans localised to the hippocampus, segmented between left and right hippocampi. Feature extraction is first performed by 2 separately trained, unsupervised based AutoEncoders, where hippocampi are fed into their respective AutoEncoder. Classification then pair of classifiers on encoded data from which each aggregated together using soft voting ensemble process. The best averaged model results recorded was with Gaussian Naïve Bayes classifier sensitivity/specificity achieved were 80%/81% respectively balanced accuracy score 80%.