基于小样本的集成学习研究

Research of ensemble learning based on small samples

为了提高集成学习在小数据量的有标记样本问题中的性能,结合半监督学习和选择性集成学习,提出了一种基于半监督回归的选择性集成算法SSRES。一方面,充分利用大量的未标记样本来辅助有标记样本的学习;另一方面,使用选择性集成学习进一步提高学习系统的泛化能力。实验结果表明,SSRES算法能够利用未标记样本和选择性集成技术提高学习器的性能。

In order to improve the performance of ensemble learning in a few labeled training examples, combining with semi-supervised learning and selective integration of learning, a new selective integration of algorithm based on semi-supervised regression namely SSRES was proposed. On the one hand, the method of a large number of unlabeled examples was used to reduce the requirement of labeled examples. On the other hand, selective integration learn was used to further improve the generalization ability of learning systems. Experiment results show that SSRES algorithm can improve the performance of learners with unlabeled example and selective integration of technology.