Classification by Clustering: Saliva MS-SELDI Data for Alzheimer's Disease Diagnosis

Sundry studies dedicated to understanding the various aspects of Alzheimer's disease exists in literature, but the unfortunate reality today, is that there still exists no cure or diagnostic tools for the disease. Early diagnosis (or at worst mild cognitive state) increases the chances of treatment. Aided with the outputs of mass spectrometry technology, this study is hinged on protein expressions in saliva (also applicable to other protein source samples) details out a novel classification model enshrined in eigenmatrix trans-formation/projection mechanism. The transformation correctly clusters elements of Mass Spectra data based on their disease label/stage. The tightness of obtained clusters is used to validate the classification of subsequent test data points, we also supported this phenomenon using the properties of Rayleighs number.