Feature selection and multi-kernel learning for sparse representation on a manifold

Wang, Jim Jing-Yan; Bensmail, Halima; Gao, Xin

doi:10.1016/j.neunet.2013.11.009

Cited by 55 publications

(22 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To some extent we alleviate this computational issue by using a joint feature weights learn-90 ing strategy [25], [26]. Let X = [x 1 , ..., x n ] ∈ R d×n be the training data matrix in the original feature space.…”

Section: Basic Idea and Formulationmentioning

confidence: 99%

Sparse preserving feature weights learning

2016

View full text Add to dashboard Cite

In this paper, we propose a novel unsupervised feature selection algorithm, named sparse preserving feature weights learning (SPFW), which is based on the recent local data representation theory, sparse representation. SPFW differs from traditional feature selection algorithms in two aspects: (1) SPFW is designed on the locality measurement criterion with sparse reconstruction residual minimization. It adaptively determines the locality based on sparse representation, instead of fixing the k-nearest neighbors in the original feature space. (2) SPFW selects the most discriminative feature subset from the whole feature set in batch mode, instead of selecting features individually. To optimize the proposed formulation, we propose an efficient iterative algorithm, where each iteration reduces to a subproblem which can be solved with some off-the-shelf toolboxes. We conduct experiments on two face datasets to evaluate the performance of feature selection in terms of classification and clustering, which demonstrate the effectiveness of the proposed algorithm.

show abstract

Section: Basic Idea and Formulationmentioning

confidence: 99%

Sparse preserving feature weights learning

2016

View full text Add to dashboard Cite

show abstract

“…(4), the optimal size of selected feature subset, i.e., d ′ , cannot be automatically determined. Moreover, it is difficult to solve this typical integer programming problem due to the 0-1 constraint on the indicator vector w. We reformulate the problem (4) by relaxing this binary constraint to be within the range of [0, 1] (Liu et al, 2011;Wang, Bensmail, & Gaoa, 2014), and obtain the final formulation of Locality Preserving Score (LPS) in the following form:…”

Section: Ideas and Algorithmmentioning

confidence: 99%

Locality preserving score for joint feature weights learning

Yan

Yang

2015

Neural Networks

View full text Add to dashboard Cite

“…Feature selection is an important issue in classification [30]. These methods can be divided in two main approaches: SIViP wrappers and filters [14].…”

Section: Feature Selectionmentioning

confidence: 99%