Feature Selection for SVMs

By Jason Aaron Edward Weston, Sayan Mukherjee, Olivier Chapelle, Massimiliano Pontil, Tomaso A. Poggio, Vladimir Naumovich Vapnik. Neural Information Processing Systems, 2000

1,183 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/46ba0c8afc9339221076f2b2e497e2a6d9ce6248

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Robust support vector method for hyperspectral data classification and knowledge discovery (2004, background): “This has sometimes led to the idea that a feature selection is not necessary when working with SVMs, which is not completely true, as shown in [20] and [50].”
Locality Preserving Projections (2003, background): “Most of works concerning feature selection have been carried out under the supervised learning paradigm [28][75][78], paying little attention to unsupervised learning [24] [60].”
Iterative RELIEF for Feature Weighting: Algorithms, Theories, and Applications (2007, unknown): “E-mail: sun@dsp.ufl.edu .”
Feature Selection via Dependence Maximization (2012, background): “Examples of criteria that satisfy both conditions include the leave-one-out error bound of SVM (Weston et al., 2000) and the mutual information (Zaffalon and Hutter, 2002).”
Use of the Zero-Norm with Linear Models and Kernel Methods (2003, methodology): “For the problem of interest, it takes the following form (see Appendix A of Weston et al. 2001a for more details about the derivation of this procedure): The method we have just defined is thus an Approximation of the zero-norm Minimization (AROM).”
Prediction Modeling Using EHR Data: Challenges, Strategies, and a Comparison of Machine Learning Approaches (2010, background): “10 Weston et al. 10 conducted an empirical study showing a number of synthetic examples where natural direct application of SVM only achieves a consistent test accuracy of 0.5 as a function of increasing training data size.”
Image retrieval: Ideas, influences, and trends of the new age (2008, methodology): “To avoid a combinatorial search, an automatic feature subset selection algorithm for SVMs is proposed in [Weston et al. 2000].”
Discriminative Semi-Supervised Feature Selection Via Manifold Regularization (2009, background): “A number of performance criteria have been proposed for feature selection, including mutual information [19], [31], maximum margin [22], [52], kernel alignment [15], [37], graph-spectrum based measures [53], [66], construction errors in neural network [35], [42], worst case probability [4], [61], an”
From Lasso regression to Feature vector machine (2005, background): “[Weston et al., 2000], [Canu et al., 2002] and [Krishnapuram et al., 2003] addressed the limitation from a different angle.”
A GA-based feature selection and parameters optimizationfor support vector machines (2006, methodology): “In the literature, only a few algorithms have been proposed for SVM feature selection (Bradley, Mangasarian, & Street, 1998; Bradley and Mangasarian, 1998; Weston et al., 2001; Guyon, Weston, Barnhill, & Bapnik, 2002; Mao, 2004).”
Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning (2002, background): “184585 712829 AA280651 1 TTG-2=Rhombotin-2=translocated in t(11;14)(p13;q11) T cell acute lymphocytic leukemia=cysteine rich protein with LIM motif”
Content-based image retrieval: approaches and trends of the new age (2005, unknown): “To avoid a combinatorial search, an automatic feature subset selection algorithm for SVMs has been proposed in [107].”
Combining SVMs with Various Feature Selection Strategies (2006, background): “[39, 7] minimize generalization bounds with respect to feature weights.”
Nightmare at test time: robust learning by feature deletion (2006, background): “The problem may be approximated via diﬁerent relaxations as in (Gilad-Bachrach et al., 2004; Weston et al., 2000).”
MIXED-INTEGER SUPPORT VECTOR MACHINE (2009, methodology): “…standard SVM method [1], Recursive Feature Elimination (RFE) method [4], L1-norm SVM method, FSV method [5], and Weston’s feature selection method [12], demonstrates either sparser solutions with roughly identical classiﬁcation performance, or an increase in classiﬁcation performance with similar…”