摘要
利用流形正则化的思想,围绕半监督学习,提出了一种针对回归问题的新算法。该算法基于流形上的正则化项和传统的正则化项相结合的方法,利用支持向量机回归已有的结果,解决半监督学习的回归问题,提高了泛化能力。通过数值试验,验证了该算法具有较好的泛化能力,对噪音具有较强的鲁棒性,与支持向量回归相比,具有更高的学习精度。
Based on the theory of manifold regularization, a new algorithm about semi-supervised learning for the problem of regression was proposed. The algorithm was deduced by the connection between the regularization term on the manifold and the classical regularization term. Using the result of support vector regression, the algorithm not only solves the problem about semi- supervised learning but also improves generalization capability. Numerical experimental results show that the algorithm enhances generalization capability and is strongly robust to noise, and has higher learning precision compared to support vector regression.
出处
《计算机应用》
CSCD
北大核心
2007年第8期1955-1958,共4页
journal of Computer Applications
关键词
半监督学习
流形正则化
支持向量回归
semi-suporvised learning
manifold regularlzation
support vector regression
