ADVANCED MACHINE LEARNING TECHNIQUES FOR ENHANCED PREDICTION OF ALZHEIMER'S DISEASE STAGES

Mohamed Douache, Ronkin Mikhail

This paper explores the use of deep learning and machine learning models to predict Alzheimer's disease stages, aiming to improve accuracy and diagnostic quality. We compared three CNN models, K-Nearest Neighbors (KNN), Temporal Convolutional Networks (TCN), and XGBoost for classifying brain scan images of different Alzheimer’s stages. CNN models were optimized for spatial feature extraction, KNN was used for instance classification, TCN captured temporal patterns, and XGBoost enhanced prediction performance through ensemble methods. Our goal was to identify the most accurate and computationally efficient model for clinical applications. Experimental results highlight the strengths and weaknesses of each approach, helping to determine the best algorithms for reliable Alzheimer’s detection.

Keywords:Alzheimer's disease, CNN, K-Nearest Neighbors (KNN), Temporal Convolutional Networks (TCN), XGBoost, machine learning, deep learning, brain scan classification, medical image analysis, prediction accuracy, ensemble learning, diagnostic tools