Abstract. When we apply support vector machines (SVM) to multiclass classification, some methods of combining the results of independent SVM for each class haven been used, However, the conventional methods may deteriorates generalization performance when the number of data in each class is small. To solve this problem, we proposed a new method, which uses only one SVM and train it to find some similarity measure between data samples. Through an experiment using real data, we confirm that the proposed method can give better classification performance than the conventional one.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.