Performance Evaluation Of Stacked Ensemble Transferred Neural Network Model Using Adni And Oasis Dataset

Abstract

Alzheimer’s disease (AD) is the most prevalent type of dementia of the nervous system that causes many brain functions to weaken e.g. memory loss. Early diagnosis of Alzheimer's disease has been shown to improve patient outcomes. Machine learning techniques that utilize magnetic resonance imaging (MRI) have been used for Alzheimer's disease diagnosis, but traditional methods require manual feature extraction by an expert, which can be complex. To address this problem, our study proposes a new ensemble learning approach using a stacked ensemble model of pre-trained convolutional neural networks as base learners and logistic regression as meta learner called Stacked Ensemble Transferred Neural Network (SETNN) model for classification of MRI images to identify Alzheimer's disease. The efficiency of the SETNN model, compared to conventional Softmax and support vector machine (SVM) methods, was evaluated using various metrics like confusion matrix, precision, accuracy and other. The suggested SETNN model performed better than other modern algorithms according to the results by achieving an accuracy of 96% when using the MRI images from OASIS dataset and achieved accuracy of 94% for ADNI dataset.