计算机科学 ›› 2014, Vol. 41 ›› Issue (4): 280-282.

• 人工智能 • 上一篇    下一篇

基于图的半监督降维算法

杨格兰,金辉霞,孟令中,朱幸辉   

  1. 湖南城市学院信息科学与工程学院 益阳413000;湖南城市学院通信与电子工程学院 益阳413000;中国科学院软件研究所基础软件测评实验室 北京100190;湖南农业大学信息科学工程学院 长沙410128
  • 出版日期:2018-11-14 发布日期:2018-11-14
  • 基金资助:
    本文受国家科技支撑计划课题(2012BAD35B07),湖南省教育厅优秀青年项目(12B023)资助

Graph-based Semi-supervised Dimensionality Reduction Algorithm

YANG Ge-lan,JIN Hui-xia,MENG Ling-zhong and ZHU Xing-hui   

  • Online:2018-11-14 Published:2018-11-14

摘要: 非线性降维和半监督学习都是近年来机器学习的热点。将半监督的方法运用到非线性降维中,提出了基于图的半监督降维的算法。该算法用等式融合的方法推出了标记传播算法的另一种表达形式,用标记传播的结果作为初始的数据映射,然后在图谱张成的线性空间中寻找最逼近初始映射的数据作为最后的半监督降维的结果。实验表明,所提算法可以获得平滑的数据映射,更接近于理想的降维效果。与标记传播算法、图谱逼近算法、无监督的降维算法的比较也体现出本算法的优越性。

关键词: 半监督学习,流形学习,标记传播,图谱理论

Abstract: Nonlinear dimensionality reduction and semi-supervised learning are both hot issues in machine learning area.Based on semi-supervised method,the article solved nonlinear dimensionality reduction problem to make up for the shortfall of ordinary methods.By using integration of equalities,a novel expression of label propagation algorithm was proposed.We used the label propagation result as the initial value mapping,and then found the best approximation to it in the graph spectral space.The experiment shows that our semi-supervised dimensionality reduction method can achieve smooth data mapping that is closer to the ideal effect.

Key words: Semi-supervised learning,Manifold learning,Label propagation,Spectral graph theory

[1] van der M L J P,Postma E O,van den Herik H J.Dimensionality reduction:A comparative review[J].Journal of Machine Learning Research,2007(1)
[2] Tenenbaum J B,de S V,Langford J C.A global geometric fra-mework for nonlinear dimensionality reduction [J].Science,290,2000:2319-2323
[3] Roweis S T,Saul L K.Nonlinear dimensionality reduction by locally linear embedding [J].Science,2000,290:2323-2326
[4] Belkin M,Niyogi P.Laplacian eigenmaps for dimensionality reduction and data representation [J].Neural Computation,2003,15(1):1373-1396
[5] Brans M M.Charting a manifold[C]∥Neural Information Proceeding Systems:Natural and Synthetic.Vancouver,Canada,2000:232-245
[6] Zhang Zhen-yue,Zha Hong-yuan.Linear low-rank approximations and nonlinear dimensionality reduction [J].Science in China Series A-Mathematics,2005,35(3):273-285
[7] Yan S C,Xu D,Zhang B,et al.Graph embedding and exten-sions:A general framework for dimensionality reduction [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2007,29(1):40-51
[8] Zhu Xiao-jin,Ghahramani Z.Learning from labeled and unlabeled data with label propagation[R].Technical Report 02-107,CMU-CALD.USA:Carnegie Mellon University,2002
[9] Zhu Xiao-jin,Lafferty J,Ghahramani Z.Semi-Supervised Lear-ning:From Gaussian Fields to Gaussian Processes[R].CMU.Technical Report,CMU-CS-03-175.USA:Carnegie Mellon University,2003
[10] Pothen,Alex,Fan C-J.Computing the Block Triangular Form of a Sparse Matrix [J].ACM Transactions on Mathematical Software,1990,16(4):303-324
[11] 李岩波,宋琼,郭新辰.基于流形距离的人工免疫半监督聚类算法[J].计算机科学,2012,9(11):204-207
[12] 刘志勇,袁媛.基于测地距离的半监督增强[J].计算机工程与应用,2011,7(21):202-204
[13] 任剑锋,梁雪,李淑红.基于非线性流形学习和支持向量机的文本分类算法[J].计算机科学,2012,9(1):261-263
[14] 罗磊,李跃华.基于LLE的分类算法及其在被动毫米波目标识别中的应用[J].电子与信息学报,2010,2(6):1306-1310
[15] 王越,王泉,吕奇峰,等.基于初始聚类中心优化和维间加权的改进k-means算法[J].重庆理工大学学报:自然科学版,2013,7(4):77-80

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!