Joint Sparse Locality Preserving Projections

作者: Haibiao Liu , Zhihui Lai , Yudong Chen

DOI: 10.1007/978-3-319-73830-7_13

关键词:

摘要: Manifold learning and feature selection have been widely studied in face recognition the past two decades. This paper focuses on making use of manifold structure datasets for extraction selection. We propose a novel method called Joint Sparse Locality Preserving Projections (JSLPP). In order to preserve datasets, we first manifold-based regression model by using nearest-neighbor graph, then \( L_{2,1} \)-norm regularization term is imposed perform At last, an efficient iterative algorithm designed solve sparse model. The convergence analysis computational complexity are presented. Experimental results indicate that JSLPP outperforms six classical state-of-the-art dimensionality reduction algorithms.

参考文章(1)
Hui Zou, Trevor Hastie, Robert Tibshirani, Sparse Principal Component Analysis Journal of Computational and Graphical Statistics. ,vol. 15, pp. 265- 286 ,(2006) , 10.1198/106186006X113430