Exploring the Efficacy of LightGBM, AdaBoost, and CatBoost in Alzheimer’s Disease Classification

Abstract

Listen

Translate article

Alzheimer’s disease (AD) is a major public health concern that requires prompt and accurate diagnosis to intervene effectively. We present a comprehensive machine learning system in our work that is specifically designed to classify AD. It incorporates various neuroimaging and clinical features, including air time1, disp index1, gmrt in air1, max x extension1, and max y extension1. We can decipher complex data correlations suggesting AD pathology using rigorous preprocessing and visualization methods, such as correlation heatmaps and 3D scatter plots. We can distinguish minute changes between AD, moderate cognitive impairment, and healthy controls using CatBoost, LightGBM, and AdaBoost classifiers. A thorough assessment of the model’s performance is provided by the rigorous evaluation metrics of accuracy, precision, recall, and F1-score, which are supplemented by detailed classification reports and confusion matrices. Learning curves also provide information about the generalization and flexibility of models. Our findings support integrated analysis approaches across many data modalities and highlight the revolutionary potential of machine learning in AD diagnosis. Developing customized treatment plans and cutting-edge clinical decision support systems is expected to improve patient outcomes and the standard of care for neurodegenerative diseases.