基于半监督回归的选择性集成算法

为了提高小数据量的有标记样本问题中学习器的性能,结合半监督学习和选择性集成学习,提出了基于半监督回归的选择性集成算法SSRES。算法基于半监督学习的基本思想,同时使用有标记样本和未标记样本训练学习器从而减少对有标记样本的需求,使用选择性集成算法GRES对不同学习器进行适当的选择,并将选择的结果结合提高学习器的泛化能力。实验结果表明,在小数据量的有标记样本问题中,该算法能够有效地提高学习器的性能。

An Ensemble Selection Algorithm Based on Semi-Supervised Regression

To improve the performance of learners in the fields without a large number of labeled training examples,by combining ensemble selection with semi-supervised learning,a new ensemble selection algorithm based on semi-supervised regression is proposed in the paper.Based on the theory of semi-supervised learning,the method uses both the labeled and unlabeled sample to train learners,and select the learners by GRES in order to improve the generalization ability of learners.Experiment results show that the propo...