论文信息 - Iterative Laplacian Score for Feature Selection

Iterative Laplacian Score for Feature Selection

Laplacian Score (LS) is a popular feature ranking based feature selection method both supervised and unsupervised. In this paper, we propose an improved LS method called Iterative Laplacian Score (IterativeLS), based on iteratively updating the nearest neighborhood graph for evaluating the importance of a feature by its locality preserving ability. Compared with LS, the key idea of IterativeLS is to gradually improve the nearest neighbor graph by discarding the least relevant features at each iteration. Experimental results on several high dimensional data sets demonstrate the effectiveness of our proposed method.

[1] Jin Xu,et al. Dictionary Learning Based on Laplacian Score in Sparse Coding , 2011, MLDM.

[2] Xiaofei He,et al. Locality Preserving Projections , 2003, NIPS.

[3] Carla E. Brodley,et al. Unsupervised Feature Selection Applied to Content-Based Retrieval of Lung Images , 2003, IEEE Trans. Pattern Anal. Mach. Intell..

[4] Deng Cai,et al. Laplacian Score for Feature Selection , 2005, NIPS.

[5] Xin Liu,et al. Document clustering based on non-negative matrix factorization , 2003, SIGIR.

[6] Mikhail Belkin,et al. Laplacian Eigenmaps and Spectral Techniques for Embedding and Clustering , 2001, NIPS.

[7] Heekuck Oh,et al. Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[8] Carla E. Brodley,et al. Feature Selection for Unsupervised Learning , 2004, J. Mach. Learn. Res..

[9] Sinisa Todorovic,et al. Local-Learning-Based Feature Selection for High-Dimensional Data Analysis , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.