Using SVM for Alzheimer’s Disease detection from 3D T1MRI
Rashmi Kumari, Shivani Goel, Subhranil Das
Abstract
The foremost cause of dementia is Alzheimer Disease (AD), where serious socioeconomic problems and health issues are concerned. AD causes adequate changes in the structure in the brain, which leads to fluctuations in behavioral patterns, psychological activities, and memory reduction. Researchers developed numerous Machine Learning (ML) algorithms for the classification of AD, Mild Cognitive Impairment (MCI), and Normal Control (NC). Early detection of AD could mitigate various risk factors. In this paper, a new ML algorithm, Hyperparameter Tuning-Twin Support Vector Machine (HPT-TSVM), has been proposed for classification. In our study, neuro-psychological data and 3D T1 weighted MRI images were considered for 202 subjects acquired from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset for validating the proposed algorithm. The proposed HPT-TSVM algorithm produces the best accuracy, sensitivity, and specificity values compared to four state-of-the-art techniques.