MRI And Environmental Data Fusion For Accurate Alzheimer’s Stage Identification Via Advanced Learning Models

Abstract

Alzheimer’s disease is a progressive brain disorder that leads to a decline in cognitive functions. Although its progression cannot be reversed once initiated, early prediction offers an opportunity to manage and slow down the disease by targeting specific protein functions associated with its development. Despite numerous efforts using machine learning techniques for early-stage diagnosis, many studies have faced challenges in achieving reliable and accurate classification results. The efficiency of transfer learning techniques for classifying the different stages of AD is examined in this case using the ensemble stacking approach where diverse transfer learning exits. Applying the Markov random field approach to the brain tissues has an impact on AD instead of feed image extraction. The brain tissues that have been harvested are used to train the base models. The second level classification meta-model is then trained by combining the predictions of the base models. The suggested model was successful in achieving a 96% accuracy rate for disease early detection.