We proposed an application to identify Alzheimer's illness in our research. The frontal section was utilized to extract the Hippocampus (H), the Sagittal section was used to analyze the Corpus Callosum (CC), and the axial section was used to work with the various aspects of the Cortex (C).Traditional machine learning approaches have relatively lower performance with larger amounts of input data. It can be challenging to detect brain abnormalities correctly and to find a solution for the automatic segmentation of brain structures. Such challenges mainly arise from the changes in settings for the acquisition of MRI scans, fluctuations in the appearance of pathology,normal anatomical variations in brain morphology,and imperfections in image acquisition. These limitations of traditional methods can be overcome by machinelearning-based methods. Moreover, machine learning can also be used to perform quantitative analysis of brain MRI through the self-learning of features, by which new features can be recognized. One of the procedures that are most commonly used in extracting features of Alzheimer’s disease from a person’s brain is called the Support Vector Machine-based Recursive Feature Elimination(SVMRFE) procedure.The specialty of SVMRFEis that it usesone slackvariable, whereas Universal Support Vector Machine (USVM) uses two slack variables.In the proposed method of classification, Global Support Vector Machine (GSVM), three slack variables are used. With the proposed method, theaccuracy is increased by 1.92%, sensitivityis increased by 10.24%in the diagnosis of Alzheimer’s disease.

GSVM, MRI, SVM,USVM, MORPHOMETRY,ALZHEIMER'S DISEASE.