A.A. Dmitriev, S.V. Kucheryavsky
Multivariate Data Analysis to Diagnose the Alzheimer's Disease using Computerized Tomographic Imaging
The present work proposes a new method of analysis of MRI tomograms typical to patients with early stages of Alzheimer's disease. The method is based on soft modeling approach, which allows us to define latent variables in initial space of features correlated with unknown quantity.
The prior study of changes in a brain structure during disease development allowed us to determine features relevant to disease diagnostics. Obtained results are used for constructing classification models which can define whether the patient is sick or not.
Key words: Alzheimer's disease, image treatment and analysis, formal modeling, multivariate data analysis