Advancing Early Diagnostic Accuracy for Alzheimer's Disease Through the Integration of Machine and Deep Learning Paradigms by Applying Multisource Datasets

Abstract

Alzheimer's disease (AD) is a progressive neurodegenerative disorder with a critical need for early and accurate diagnosis. In this research, various machine learning (ML) and deep learning (DL) techniques are employed for identifying reliable biomarkers for early-stage AD detection. This research paper presents a comprehensive approach utilizing a myriad of artificial intelligence (AI) and machine learning (ML), i.e., self-attention mechanisms with Convolutional Neural Networks (CNN) model has been proposed for predicting the onset of AD.  The study has compared these models including Gaussian Naive Bayes (GNB), Decision Trees (DT), Random Forest (RF), XGBoost, Voting Classifier (VC) where the proposed Feature based CNN (F-CNN) network produces the highest classification accuracy, sensitivity, recall, AUC and F1 scores by utilizing datasets such as the Alzheimer's Disease Neuroimaging Initiative (ADNI). The proposed approach holds promise for early detection and intervention in AD, potentially enabling clinicians to intervene before the onset of clinical symptoms. The findings of this research could significantly contribute to improving patient outcomes and advancing our understanding of Alzheimer's disease pathology.