Advanced MRI-Based Alzheimer’s Disease Classification with Hybrid Convolutional Neural Networks

The key to effective therapy and management of Alzheimer's disease (AD), a progressive neurological condition, is obtaining a prompt and precise diagnosis. Support Vector Machines (SVMs), Random Forest (RFs), and Gradient Boosting Machines (GBMs) are some of the most used traditional machine learning algorithms for Alzheimer's disease (AD) classification, but they struggle to deal with the complexity of medical imaging data. On the other hand, Convolutional Neural Networks (CNNs) are great at image classification and have shown impressive results in analyzing MRI scans for AD detection. Plus, combining CNNs with traditional machine learning techniques can enhance both accuracy and reliability. This study explores hybrid methodologies that integrate Convolutional Neural Networks (CNNs) with conventional machine learning techniques, such as the combination of CNN with Support Vector Machines (SVM), CNN with Gradient Boosting Machines (GBM), and CNN with Random Forest (RF).The study achieved notable classification performance, with the CNN-SVM model reaching an accuracy of 97% and the CNN-RF model achieving 83%. These results indicate that blending deep learning with conventional machine learning methods can enhance the ability of Alzheimer’s disease (AD) diagnosis, promoting more reliable early detection and supporting the creation of effective intervention strategies.