ICGA-ELM classifier for Alzheimer's disease detection
B. S. Mahanand, S. Suresh, N. Sundararajan, M. Aswatha Kumar · 2013
In this paper, we present an approach for Alzheimer's disease detection using voxel-based morphometric features and an extreme learning machine classifier. For feature selection, Integer Coded Genetic Algorithm along with the Extreme Learning Machine classifier (referred to here as the ICGA-ELM classifier) is proposed. The ICGA-ELM classifier is used to select the best set of features (highest classification accuracy) obtained from the voxel-based morphometry analysis. In our study, Open Access Series of Imaging Studies (OASIS) data set is used to evaluate the performance of the proposed ICGA-ELM classifier. The results of the ICGA-ELM classifier is compared with that of the Support Vector Machine (SVM) classifier. The results indicate that the ICGA-ELM classifier produces a mean testing accuracy of 91.86% with only 10 features whereas, the SVM produces a mean testing accuracy of 86.84% for the same set. The ICGA selected features are also mapped back into the standard brain space to identify the regions likely to onset of Alzheimer's disease.