Ranking a Random Feature for Variable and Feature Selection

By Hervé Stoppiglia, Gérard Dreyfus, Rémi Dubois, Yacine Oussar. Journal of machine learning research, 2003

351 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/bf48afa4422fc02c0d07bbbf262cb8254a365359

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

An improved approximation algorithm for the column subset selection problem (2008, result): “We should note that similar formulations appeared in [23, 33, 36, 37, 26, 1].”
Feature Selection with the Boruta Package (2010, methodology): “The algorithm is an extension of the idea introduced by Stoppiglia, Dreyfus, Dubois, and Oussar (2003) to determine relevance by comparing the relevance of the real features to that of the random probes.”
Unsupervised feature selection for principal components analysis (2008, background): “More recently, [39, 42, 43, 28] present algorithms based on spectral methods, that, to the best of our understanding, do not come with provable guarantees of the type that we seek here.”
Causality : Objectives and Assessment (2010, background): “, 2006) or “probes” (Stoppiglia et al., 2003; Guyon and Dreyfus, 2006), which are, by construction, not predictive of a target variable of interest.”
An introduction to variable and feature selection (2003, methodology): “GramSchmidt orthogonalization is presented in this issue as an alternative to OBD (Stoppiglia et al., 2003).”
A survey on feature selection methods (2014, methodology): “In [27] a random variable called probe is used to rank the features using Gram-Schmidt orthogonalization.”
Feature Selection with Ensembles, Artificial Variables, and Redundancy Elimination (2009, background): “In practice, most alg rithms just try to remove irrelevant features and then apply some heuristics that remove “possibly”redundant variables.”