In this paper, we focus on the classification problem to semisupervised learning. Semi-supervised learning is a learning task from both labeled and unlabeled data examples. We propose a novel semisupervised learning algorithm using a self-training framework and support vector machine. Self-training is one of the wrapper-based semisupervised algorithms in which the base classifier assigns labels to unlabeled data at each iteration and the classifier re-train on a larger training set at the next training step. However, the performance of this algorithm strongly depends on the selected newly-labeled examples. In this paper, a novel self-training algorithm is proposed, which improves the learning performance using the idea of the Apollonius circle to find neighborhood examples. The proposed algorithm exploits a geometric structure to optimize the self-training process. The experimental results demonstrate that the proposed algorithm can effectively improve the performance of the constructed classification model.
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.