A robust semi-supervised boosting method using linear programming

Shaodan Zhai, Tian Xia, Ming Tan, Shaojun Wang

Abstract: We propose a novel semi-supervised boosting algorithm using linear programming, which explicitly maximizes the margin over both labeled and unlabeled data. Experiments conducted on a number of UCI datasets and synthetic data show that, the algorithm we propose performs better than the state-of-the-art supervised and semi-supervised boosting algorithms, and it is more robust with noisy data.