Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Neighbourhood discernibility degree-based semisupervised feature selection for partially labelled mixed-type data with granular ball

Neighbourhood discernibility degree-based semisupervised feature selection for partially labelled... Feature selection can effectively decrease data dimensions by selecting a relevant feature subset. Rough set theory provides a powerful theoretical framework for the feature selection of categorical data with complete labels. However, in reality, the given datasets have only a small number of objects with label information and many unlabelled objects. Furthermore, most of feature selection approaches are computationally expensive. To address the above problems, a semisupervised feature selection algorithm based on neighbourhood discernibility with pseudolabelled granular balls is proposed. First, the set of granular balls based on the purity is generated, which reduces the universe space by sampling. Then, the neighbourhood discernibility is proposed to validate the importance of the candidate features for both labelled and unlabelled objects. Finally, an ensemble voting algorithm is designed to execute feature selection, and a feature subset with satisfactory performance is selected fairly not arbitrarily. On UCI datasets, experimental results verify the advantage of the proposed feature selection algorithm in terms of the feature subset size, classification accuracy and computational time against other algorithms. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Intelligence Springer Journals

Neighbourhood discernibility degree-based semisupervised feature selection for partially labelled mixed-type data with granular ball

Shu, Wenhao; Yu, Jianhui; Chen, Ting; Qian, Wenbin
Applied Intelligence , Volume OnlineFirst – Jun 29, 2023
Download PDF
21 pages

Loading next page...
 
/lp/springer-journals/neighbourhood-discernibility-degree-based-semisupervised-feature-HPQWx7LD6q

References (47)

Publisher
Springer Journals
Copyright
Copyright © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
ISSN
0924-669X
eISSN
1573-7497
DOI
10.1007/s10489-023-04657-7
Publisher site
See Article on Publisher Site

Abstract

Feature selection can effectively decrease data dimensions by selecting a relevant feature subset. Rough set theory provides a powerful theoretical framework for the feature selection of categorical data with complete labels. However, in reality, the given datasets have only a small number of objects with label information and many unlabelled objects. Furthermore, most of feature selection approaches are computationally expensive. To address the above problems, a semisupervised feature selection algorithm based on neighbourhood discernibility with pseudolabelled granular balls is proposed. First, the set of granular balls based on the purity is generated, which reduces the universe space by sampling. Then, the neighbourhood discernibility is proposed to validate the importance of the candidate features for both labelled and unlabelled objects. Finally, an ensemble voting algorithm is designed to execute feature selection, and a feature subset with satisfactory performance is selected fairly not arbitrarily. On UCI datasets, experimental results verify the advantage of the proposed feature selection algorithm in terms of the feature subset size, classification accuracy and computational time against other algorithms.

Journal

Applied Intelligence – Springer Journals

Published: Jun 29, 2023

Keywords: Feature selection; Granular ball; Mixed-type data; Semi-supervised; Ensemble selector

There are no references for this article.