A Cognitive Ensemble Classifier based on Risk Sensitive Hinge Loss Function for Alzheimer’s Disease diagnosis in early stages

In this paper, we present a novel approach for Alzheimer's disease diagnosis based on magnetic resonance images (MRI) using cognitive ensemble classifier. The proposed cognitive ensemble classifier employs a binary coded genetic algorithm (BCGA) for feature selection and an extreme learning machine (ELM) for classification. The cognitive ensemble classifier uses voxel based morphometric features extracted from MRI and subset of features are selected using BCGA. These selected features are then used to train ELM classifiers and an ensemble classifier is built based on risk sensitive hinge loss function. The study has been conducted using the open access series of imaging studies dataset and Alzheimer's disease neuroimaging initiative dataset. The results clearly indicate that the proposed cognitive ensemble classifier produces better accuracy for AD diagnosis compared to other similar methods.