Match!

On Similarity Preserving Feature Selection

Published on Mar 1, 2013in IEEE Transactions on Knowledge and Data Engineering3.86
· DOI :10.1109/TKDE.2011.222
Zheng Zhao24
Estimated H-index: 24
(SAS: SAS Institute),
Lei Wang38
Estimated H-index: 38
(UOW: University of Wollongong)
+ 1 AuthorsJieping Ye57
Estimated H-index: 57
(ASU: Arizona State University)
Abstract
In the literature of feature selection, different criteria have been proposed to evaluate the goodness of features. In our investigation, we notice that a number of existing selection criteria implicitly select features that preserve sample similarity, and can be unified under a common framework. We further point out that any feature selection criteria covered by this framework cannot handle redundant features, a common drawback of these criteria. Motivated by these observations, we propose a new "Similarity Preserving Feature Selection” framework in an explicit and rigorous way. We show, through theoretical analysis, that the proposed framework not only encompasses many widely used feature selection criteria, but also naturally overcomes their common weakness in handling feature redundancy. In developing this new framework, we begin with a conventional combinatorial optimization formulation for similarity preserving feature selection, then extend it with a sparse multiple-output regression formulation to improve its efficiency and effectiveness. A set of three algorithms are devised to efficiently solve the proposed formulations, each of which has its own advantages in terms of computational complexity and selection performance. As exhibited by our extensive experimental study, the proposed framework achieves superior feature selection performance and attractive properties.
  • References (49)
  • Citations (191)
References49
Newest
#1Le Song (Georgia Institute of Technology)H-Index: 47
#2Alexander J. Smola (Yahoo!)H-Index: 97
Last.Karsten M. Borgwardt (MPG: Max Planck Society)H-Index: 44
view all 5 authors...
#1Luping Zhou (ANU: Australian National University)H-Index: 15
#2Lei Wang (ANU: Australian National University)H-Index: 38
Last.Chunhua Shen (NICTA)H-Index: 54
view all 3 authors...
Jul 11, 2009 in IJCAI (International Joint Conference on Artificial Intelligence)
#1Zenglin Xu (CUHK: The Chinese University of Hong Kong)H-Index: 20
#2Ruozhong Jin (MSU: Michigan State University)H-Index: 62
Last.Irwin King (CUHK: The Chinese University of Hong Kong)H-Index: 51
view all 4 authors...
Jun 18, 2009 in UAI (Uncertainty in Artificial Intelligence)
#1Jun Liu (ASU: Arizona State University)H-Index: 29
#2Shuiwang Ji (ASU: Arizona State University)H-Index: 32
Last.Jieping Ye (ASU: Arizona State University)H-Index: 57
view all 3 authors...
Jun 14, 2009 in ICML (International Conference on Machine Learning)
#1Blake Shaw (Columbia University)H-Index: 8
#2Tony Jebara (Columbia University)H-Index: 41
Jun 14, 2009 in ICML (International Conference on Machine Learning)
#1Liang Sun (ASU: Arizona State University)H-Index: 12
#2Shuiwang Ji (ASU: Arizona State University)H-Index: 32
Last.Jieping Ye (ASU: Arizona State University)H-Index: 57
view all 3 authors...
#1Trevor HastieH-Index: 103
#2Robert TibshiraniH-Index: 128
Last.Jerome H. FriedmanH-Index: 63
view all 3 authors...
#1Andreas Argyriou (UCL: University College London)H-Index: 14
#2Theodoros Evgeniou (Ad: INSEAD)H-Index: 25
Last.Massimiliano Pontil (UCL: University College London)H-Index: 52
view all 3 authors...
Cited By191
Newest
#1Shiping Wang (FZU: Fuzhou University)H-Index: 13
#2Jiawei Chen (FZU: Fuzhou University)
Last.Genggeng Liu (FZU: Fuzhou University)
view all 4 authors...
#1Xiaohong Han (TUT: Taiyuan University of Technology)H-Index: 5
#2Ping Liu (TUT: Taiyuan University of Technology)
Last.Dengao Li (TUT: Taiyuan University of Technology)
view all 4 authors...
#1Zhao Kang (University of Electronic Science and Technology of China)H-Index: 14
#2Honghui Xu (University of Electronic Science and Technology of China)H-Index: 1
Last.Zenglin Xu (University of Electronic Science and Technology of China)H-Index: 20
view all 5 authors...
#1Luyao Teng (VU: Victoria University, Australia)H-Index: 3
#2Zhenye Feng (GDUT: Guangdong University of Technology)
Last.Yanchun Zhang (VU: Victoria University, Australia)H-Index: 36
view all 7 authors...
#1Zhaolong Ling (Hefei University of Technology)H-Index: 1
#2Kui Yu (Hefei University of Technology)H-Index: 9
Last.Xindong WuH-Index: 51
view all 6 authors...
#1Rui Zhang (NPU: Northwestern Polytechnical University)H-Index: 7
#2Feiping Nie (NPU: Northwestern Polytechnical University)H-Index: 57
Last.Xian Wei (CAS: Chinese Academy of Sciences)H-Index: 2
view all 4 authors...
#1Dong Huang (SCAU: South China Agricultural University)H-Index: 9
#2Xiaosha Cai (SCAU: South China Agricultural University)
Last.Chang-Dong Wang (SYSU: Sun Yat-sen University)H-Index: 15
view all 3 authors...
#1Rui Ma (National University of Defense Technology)
#2Yijie Wang (National University of Defense Technology)H-Index: 2
Last.Li Cheng (National University of Defense Technology)H-Index: 1
view all 3 authors...
#1Zhaleh Manbari (UKH: University of Kurdistan Hewler)H-Index: 3
#2Fardin Akhlaghian Tab (UKH: University of Kurdistan Hewler)H-Index: 7
Last.Chiman Salavati (UKH: University of Kurdistan Hewler)H-Index: 3
view all 3 authors...
View next paperl 2,1 -norm regularized discriminative feature selection for unsupervised learning