Skip to content
conftrace_
2003 JMLR JMLR 2003

Ranking a Random Feature for Variable and Feature Selection

Abstract

We describe a feature selection method that can be applied directly to models that are linear with respect to their parameters, and indirectly to others. It is independent of the target machine. It is closely related to classical statistical hypothesis tests, but it is more intuitive, hence more suitable for use by engineers who are not statistics experts. Furthermore, some assumptions of classical tests are relaxed. The method has been used successfully in a number of applications that are briefly described. [abs] [pdf] [ps.gz] [ps]

📈 Trend Setter - Statistical Learning
🧭 Keyword Pioneer - hypothesis test
🐣 Hot Topic Early Bird - feature selection

Authors

Hervé Stoppiglia, Gérard Dreyfus, Remi Dubois, Yacine Oussar

Topics

Machine Learning > Core Methods > Classification Machine Learning > Core Methods > Metric Learning Machine Learning > Optimization & Theory > Statistical Learning Machine Learning > Core Methods > Feature Selection

Keywords

variable selection feature selection linear model hypothesis test statistical test feature ranking
Download PDF

More from JMLR 2003

All JMLR 2003 papers →
Variable Selection Using SVM-based Criteria
Overfitting in Making Comparisons Between Variable Selection Methods
MLPs (Mono-Layer Polynomials and Multi-Layer Perceptrons) for Nonlinear Modeling
Feature Extraction by Non-Parametric Mutual Information Maximization
Use of the Zero-Norm with Linear Models and Kernel Methods
On Nearest-Neighbor Error-Correcting Output Codes with Application to All-Pairs Multiclass Support Vector Machines