Comparison of Deep Learning and Support Vector Machine Learning for Subgroups of Multiple Sclerosis

Machine learning methods are frequently used for data sets in many fields including medicine for purposes of feature extraction and pattern recognition. This study includes lesion data obtained from Magnetic Resonance images taken in three different years and belonging to 120 individuals (with 76 RRMS, 6 PPMS, 38 SPMS). Many alternative methods are used nowadays to be able to find out the strong and distinctive features of Multiple Sclerosis based on MR images. Deep learning has the working capacity pertaining to a much wider scaled space (120 x 228), less dimension (50 x 228) (also referred to as distinctive) feature space and SVM (120 x 228). Deep learning has formed a more skillful system in the classification of MS subgroups by working with fewer sets of features compared to SVM algorithm. Deep learning algorithm has a better accuracy rate in comparing the MS subgroups compared to multiclass SVM algorithm kernel types which are among the conventional machine learning systems.