Greedy feature construction

We present an effective method for supervised feature construction. The main goal of the approach is to construct a feature representation for which a set of linear hypotheses is of sufficient capacity -- large enough to contain a satisfactory solution to the considered problem and small enough to a...

Full description

Bibliographic Details
Main Authors: Oglic, Dino, Gaertner, Thomas
Format: Article
Published: Curran Associates Inc. 2016
Online Access:https://eprints.nottingham.ac.uk/38608/
_version_ 1848795651400269824
author Oglic, Dino
Gaertner, Thomas
author_facet Oglic, Dino
Gaertner, Thomas
author_sort Oglic, Dino
building Nottingham Research Data Repository
collection Online Access
description We present an effective method for supervised feature construction. The main goal of the approach is to construct a feature representation for which a set of linear hypotheses is of sufficient capacity -- large enough to contain a satisfactory solution to the considered problem and small enough to allow good generalization from a small number of training examples. We achieve this goal with a greedy procedure that constructs features by empirically fitting squared error residuals. The proposed constructive procedure is consistent and can output a rich set of features. The effectiveness of the approach is evaluated empirically by fitting a linear ridge regression model in the constructed feature space and our empirical results indicate a superior performance of our approach over competing methods.
first_indexed 2025-11-14T19:35:28Z
format Article
id nottingham-38608
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:35:28Z
publishDate 2016
publisher Curran Associates Inc.
recordtype eprints
repository_type Digital Repository
spelling nottingham-386082020-05-04T18:27:31Z https://eprints.nottingham.ac.uk/38608/ Greedy feature construction Oglic, Dino Gaertner, Thomas We present an effective method for supervised feature construction. The main goal of the approach is to construct a feature representation for which a set of linear hypotheses is of sufficient capacity -- large enough to contain a satisfactory solution to the considered problem and small enough to allow good generalization from a small number of training examples. We achieve this goal with a greedy procedure that constructs features by empirically fitting squared error residuals. The proposed constructive procedure is consistent and can output a rich set of features. The effectiveness of the approach is evaluated empirically by fitting a linear ridge regression model in the constructed feature space and our empirical results indicate a superior performance of our approach over competing methods. Curran Associates Inc. 2016-12-05 Article PeerReviewed Oglic, Dino and Gaertner, Thomas (2016) Greedy feature construction. Advances in Neural Information Processing Systems, 29 . pp. 3945-3953. ISSN 1049-5258 http://papers.nips.cc/paper/6557-greedy-feature-construction
spellingShingle Oglic, Dino
Gaertner, Thomas
Greedy feature construction
title Greedy feature construction
title_full Greedy feature construction
title_fullStr Greedy feature construction
title_full_unstemmed Greedy feature construction
title_short Greedy feature construction
title_sort greedy feature construction
url https://eprints.nottingham.ac.uk/38608/
https://eprints.nottingham.ac.uk/38608/