Since 1998 there has been significant interest in supervised learning algorithms that combine labeled and unlabeled data for text learning tasks. The co-training algorithm applied to datasets which have a natural separation of their features into two disjoint sets. In this paper, we demonstrate that when learning from labeled and unlabeled data using co-training algorithm, selecting those document examples first which have two parts of best matching features can obtain a good performance. © Springer-Verlag Berlin Heidelberg 2007.
CITATION STYLE
Wang, H., Ji, L., & Zuo, W. (2007). Best-match method used in co-training algorithm. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4819 LNAI, pp. 401–409). Springer Verlag. https://doi.org/10.1007/978-3-540-77018-3_40
Mendeley helps you to discover research relevant for your work.