New Hybrid Feature Selection Approaches Based on ANN and Novel Sparsity Norm

Feature selection is crucial for minimizing redundancy in information and addressing the limitations of traditional classification methods when dealing with large datasets and numerous features in many machine learning applications. To improve the classification, this article introduced two hybrid m...

Full description

Saved in:
Bibliographic Details
Main Authors: Khadijeh Nemati, Amir Hosein Refahi Sheikhani, Sohrab Kordrostami, Kamrad Khoshhal Roudposhti
Format: Article
Language:English
Published: Wiley 2024-01-01
Series:Journal of Electrical and Computer Engineering
Online Access:http://dx.doi.org/10.1155/2024/7112770
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846166280596357120
author Khadijeh Nemati
Amir Hosein Refahi Sheikhani
Sohrab Kordrostami
Kamrad Khoshhal Roudposhti
author_facet Khadijeh Nemati
Amir Hosein Refahi Sheikhani
Sohrab Kordrostami
Kamrad Khoshhal Roudposhti
author_sort Khadijeh Nemati
collection DOAJ
description Feature selection is crucial for minimizing redundancy in information and addressing the limitations of traditional classification methods when dealing with large datasets and numerous features in many machine learning applications. To improve the classification, this article introduced two hybrid methods utilizing a genetic algorithm and a gray wolf algorithm with structured dispersion norms for feature selection. These techniques involved the utilization of a genetic algorithm and a gray wolf algorithm for feature selection. The features selected by these algorithms were used in the classification process by employing a two-layer perceptron as a classifier. The novel sparse norm is employed to assess and compute classification errors in these methodologies. To assess the effectiveness of the suggested techniques, they were compared with the existing feature selection methods using various publicly accessible datasets. The results of the experiments consistently demonstrate that the proposed methods outperform other approaches.
format Article
id doaj-art-912ad3518c1140cf96eed02b7761614f
institution Kabale University
issn 2090-0155
language English
publishDate 2024-01-01
publisher Wiley
record_format Article
series Journal of Electrical and Computer Engineering
spelling doaj-art-912ad3518c1140cf96eed02b7761614f2024-11-16T00:00:03ZengWileyJournal of Electrical and Computer Engineering2090-01552024-01-01202410.1155/2024/7112770New Hybrid Feature Selection Approaches Based on ANN and Novel Sparsity NormKhadijeh Nemati0Amir Hosein Refahi Sheikhani1Sohrab Kordrostami2Kamrad Khoshhal Roudposhti3Department of Applied Mathematics and Computer ScienceDepartment of Applied Mathematics and Computer ScienceDepartment of Applied Mathematics and Computer ScienceDepartment of Computer EngineeringFeature selection is crucial for minimizing redundancy in information and addressing the limitations of traditional classification methods when dealing with large datasets and numerous features in many machine learning applications. To improve the classification, this article introduced two hybrid methods utilizing a genetic algorithm and a gray wolf algorithm with structured dispersion norms for feature selection. These techniques involved the utilization of a genetic algorithm and a gray wolf algorithm for feature selection. The features selected by these algorithms were used in the classification process by employing a two-layer perceptron as a classifier. The novel sparse norm is employed to assess and compute classification errors in these methodologies. To assess the effectiveness of the suggested techniques, they were compared with the existing feature selection methods using various publicly accessible datasets. The results of the experiments consistently demonstrate that the proposed methods outperform other approaches.http://dx.doi.org/10.1155/2024/7112770
spellingShingle Khadijeh Nemati
Amir Hosein Refahi Sheikhani
Sohrab Kordrostami
Kamrad Khoshhal Roudposhti
New Hybrid Feature Selection Approaches Based on ANN and Novel Sparsity Norm
Journal of Electrical and Computer Engineering
title New Hybrid Feature Selection Approaches Based on ANN and Novel Sparsity Norm
title_full New Hybrid Feature Selection Approaches Based on ANN and Novel Sparsity Norm
title_fullStr New Hybrid Feature Selection Approaches Based on ANN and Novel Sparsity Norm
title_full_unstemmed New Hybrid Feature Selection Approaches Based on ANN and Novel Sparsity Norm
title_short New Hybrid Feature Selection Approaches Based on ANN and Novel Sparsity Norm
title_sort new hybrid feature selection approaches based on ann and novel sparsity norm
url http://dx.doi.org/10.1155/2024/7112770
work_keys_str_mv AT khadijehnemati newhybridfeatureselectionapproachesbasedonannandnovelsparsitynorm
AT amirhoseinrefahisheikhani newhybridfeatureselectionapproachesbasedonannandnovelsparsitynorm
AT sohrabkordrostami newhybridfeatureselectionapproachesbasedonannandnovelsparsitynorm
AT kamradkhoshhalroudposhti newhybridfeatureselectionapproachesbasedonannandnovelsparsitynorm