Alzheimer's Disease Detection: A Comparative Study of Machine Learning Models and Multilayer Perceptron

The intersection of Artificial Intelligence (AI) and medical science has shown great promise in recent years for addressing complex medical challenges, including the early detection of Alzheimer's disease (AD). Alzheimer's disease presents a significant challenge in healthcare, and despite advancements in medical science, a cure has yet to be found. Early detection and accurate prediction of AD progression are crucial for improving patient outcomes. This study comprehensively evaluates four Machine Learning (ML) models and one Perceptron Model for early detection of AD using the Open Access Series of Imaging Studies (OASIS) dataset. The evaluated models include Logistic Regression, Random Forest, XGBoost, CatBoost, and a Multi-layer Perceptron (MLP). This study assesses the performance of each model, on metrics like accuracy, precision, recall, and AUC ROC. The MLP model emerges as the top performer, achieving an impressive accuracy of 95 %, highlighting its efficacy in accurately predicting AD status based on biomarker indicators. While other models, such as Logistic Regression (85 %), Random Forest (87 %), XGBoost (83 %), and CatBoost (89 %), demonstrate considerable accuracy, they are outperformed by the MLP model.