Comparing Machine Learning Feature Selection Methods for Dementia Anatomical Brain MRI

There has been renewed focus on applying ML techniques to neurodegenerative disorders in recent years. Dementia is one of these new worldwide health problems, and early diagnosis of it is particularly beneficial. Alzheimer's disease (AD) is the most frequent kind of dementia. The areas of the brain that are generally affected by dementia are those that affect a person's ability to think, retain information, and communicate. An investigation of the similarities and differences between several different ML algorithms, including Convolutional Neural Networks (CNN), Random Forest, Support
Vector Machine (SVM), and others. It reports a split-half resampling examination of many data-driven feature selection and classification strategies for whole-brain voxel-based classification of Magnetic Resonance Imaging (MRI) scans. A comparative examination of all the other ML approaches reveals that the SVM methodology predicts greater accuracy (97%), higher specificity
(100%), and higher sensitivity (45%) for the diagnosis of dementia than any of the other techniques.
Keywords:Dementia diagnosis,Alzheimer's disease, Machine learning, Neurodegenerative,MRI, Feature Selection