Semi-Supervised Nonlinear Dimensionality Reduction with Pairwise Constraints

摘要：

The problem of semi-supervised dimensionality reduction with kernels called KS2 DR is considered for semi-supervised learning. In this setting, domain knowledge in the form of pair constraints is adopted to specify whether pairs of instances belong to the same class or not. KS2 DR can project the samples data onto a set of useful features and preserve the structure of unlabeled samples data as well as both similar and dissimilar constraints defined in the feature space, under which the samples with different class labels are easier to be effectively partitioned from each other. We demonstrate the practical usefulness and high scalability of KS2 DR algorithms in data visualization and classification tasks through extensive simulation studies. Experimental results show the proposed methods can almost always achieve the highest accuracy when the dimension is reduced. And KS2 DR methods outperform some established dimensionality reduction methods no matter how many numbers of constraints, dimensions are used.

关键词： Semi-supervised learning Kernel feature space Dimensionality reduction (Dis-) similar constraints

作者: Min Chen Zhao Zhang

作者单位: Department of Computer Science and Technology Hunan Institue of Technology Hunan,China School of Computer Science and Technology Nanjing Forestry University Nanjing,China

会议类型: 国际会议

会议名称: The 2nd IEEE International Conference on Advanced Computer Control(第二届先进计算机控制国际会议 ICACC 2010)

会议地点: 沈阳

会议语种:英文

页码: 116-121

在线出版日期: 2010-03-27（万方平台首次上网日期，不代表论文的发表时间）