Graph-Guided Universum Learning in Generalized Eigenvalue Proximal SVMs for Alzheimer's Disease Classification
Researchers have developed new graph-guided machine learning models, UG-GEPSVM and IUG-GEPSVM, for classifying Alzheimer's disease (AD) using structural MRI data. These models incorporate information from mild cognitive impairment (MCI) subjects by constructing a graph that captures the geometric structure of MCI samples. Experiments on the ADNI dataset show that these novel approaches outperform existing methods, with UG-GEPSVM achieving an average AUC of 88.07% and demonstrating stable performance across varying noise levels. AI
IMPACT These models offer improved accuracy for early Alzheimer's detection, potentially aiding in timely intervention and disease management.