Alzheimer’s Disease Diagnosis via Deep Factorization Machine Models

The current state-of-the-art deep neural networks (DNNs) for Alzheimer's Disease diagnosis use different biomarker combinations to classify patients, but do not allow extracting knowledge about the interactions of biomarkers. However, improve our understanding disease, it is paramount extract such from learned model. In this paper, we propose a Deep Factorization Machine model that combines ability DNNs learn complex relationships and ease interpretability linear proposed has three parts: (i) an embedding layer deal with sparse categorical data, (ii) efficiently pairwise interactions, (iii) DNN implicitly higher order interactions. experiments on data Neuroimaging Initiative, demonstrate classifies cognitive normal, mild impaired, demented patients more accurately than competing models. addition, show valuable among biomarkers can be obtained.