Margin and domain classifications for target detection over huge population of outliers

Multi-category classification is an ongoing research topic with numerous applications. In this dissertation, a novel approach called margin and domain integrated classifier (MDIC) is addressed. It merges the conventional support vector machine (SVM) and support vector domain description (SVDD) classifiers, and handles multi-class problems as a combination of several target classes plus outliers. The basic idea behind the proposed approach is that target classes possess structured characteristics while outliers scatter around in the feature space. In our approach the domain description and large margin discrimination are adjustable and therefore yield higher classification accuracy which leads to better performance than conventional classifiers. The properties of MDIC are analyzed and the performance comparisons using synthetic and real data are presented.