Decision Tree Based System for Alzheimer's Disease Classification
Theethat Meemak, Peeravit Tassiri, Tamon Santithum, Thipudon Khunprasroed, Narumol Chumuang
Abstract
Alzheimer's disease (AD) is a leading cause of dementia, underscoring the need for accurate diagnostic support systems. This study presents a machine learning framework for AD classification using a dataset of 2,149 patient records. The workflow included data cleaning, normalization, feature selection, and data balancing, followed by evaluation of multiple classifiers. Algorithms tested included Decision Tree, Decision Table, Bagging, Filtered Classifier, Attribute Selected Classifier, Iterative Classifier Optimizer, LogitBoost, and BayesNet under three training:validation:testing regimes (50:10:40, 60:10:30, 70:10:20). Results show that meta-learning and ensemble approaches, particularly Attribute Selected Classifier, Bagging, and Filtered Classifier, consistently achieved <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\sim 95\%$</tex> accuracy, outperforming traditional classifiers. These findings demonstrate the effectiveness of combining preprocessing with ensemble methods for reliable AD prediction.