DCT-Based Channel Attention for Multivariate Time Series Classification

<p dir="ltr">This article introduces a novel DCT-based channel attention (DCA) mechanism for time series classification (TSC) using convolutional neural networks (CNNs). Traditional squeeze-and-excitation (SE) mechanisms rely on global average pooling to model channel-wise interdepen...

Full description

Saved in:
Bibliographic Details
Main Author: Amine Haboub (22392097) (author)
Other Authors: Hamza Baali (14603380) (author), Abdesselam Bouzerdoum (17900021) (author)
Published: 2025
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1864513531583397888
author Amine Haboub (22392097)
author2 Hamza Baali (14603380)
Abdesselam Bouzerdoum (17900021)
author2_role author
author
author_facet Amine Haboub (22392097)
Hamza Baali (14603380)
Abdesselam Bouzerdoum (17900021)
author_role author
dc.creator.none.fl_str_mv Amine Haboub (22392097)
Hamza Baali (14603380)
Abdesselam Bouzerdoum (17900021)
dc.date.none.fl_str_mv 2025-07-25T12:00:00Z
dc.identifier.none.fl_str_mv 10.1109/ojcs.2025.3586682
dc.relation.none.fl_str_mv https://figshare.com/articles/journal_contribution/DCT-Based_Channel_Attention_for_Multivariate_Time_Series_Classification/30860261
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Information and computing sciences
Artificial intelligence
Machine learning
Channel attention
convolutional neural networks
deep learning
discrete cosine transform
time series classification
Computational modeling
Feature extraction
Vectors
Training
Standards
Adaptation models
Accuracy
dc.title.none.fl_str_mv DCT-Based Channel Attention for Multivariate Time Series Classification
dc.type.none.fl_str_mv Text
Journal contribution
info:eu-repo/semantics/publishedVersion
text
contribution to journal
description <p dir="ltr">This article introduces a novel DCT-based channel attention (DCA) mechanism for time series classification (TSC) using convolutional neural networks (CNNs). Traditional squeeze-and-excitation (SE) mechanisms rely on global average pooling to model channel-wise interdependencies, which may oversimplify complex temporal dynamics. The proposed DCA model leverages discrete cosine transform (DCT) coefficients to incorporate frequency-domain information, capturing a broader spectrum of temporal features. Two selection criteria are employed to identify the most informative DCT coefficients for constructing the attention map. The first criterion utilizes the lowest frequency coefficients, whereas the second criterion selects the coefficients exhibiting the highest energy to construct the attention map. Comprehensive experiments on twelve diverse TSC datasets demonstrate that DCA consistently outperforms state-of-the-art attention mechanisms, achieving an average improvement of 2.2% in classification accuracy.</p><h2 dir="ltr">Other Information</h2><p dir="ltr">Published in: IEEE Open Journal of the Computer Society<br>License: <a href="https://creativecommons.org/licenses/by/4.0/deed.en" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/ojcs.2025.3586682" target="_blank">https://dx.doi.org/10.1109/ojcs.2025.3586682</a></p>
eu_rights_str_mv openAccess
id Manara2_a9375749693e86cc34335ddcad7e2b51
identifier_str_mv 10.1109/ojcs.2025.3586682
network_acronym_str Manara2
network_name_str Manara2
oai_identifier_str oai:figshare.com:article/30860261
publishDate 2025
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling DCT-Based Channel Attention for Multivariate Time Series ClassificationAmine Haboub (22392097)Hamza Baali (14603380)Abdesselam Bouzerdoum (17900021)Information and computing sciencesArtificial intelligenceMachine learningChannel attentionconvolutional neural networksdeep learningdiscrete cosine transformtime series classificationComputational modelingFeature extractionVectorsTrainingStandardsAdaptation modelsAccuracy<p dir="ltr">This article introduces a novel DCT-based channel attention (DCA) mechanism for time series classification (TSC) using convolutional neural networks (CNNs). Traditional squeeze-and-excitation (SE) mechanisms rely on global average pooling to model channel-wise interdependencies, which may oversimplify complex temporal dynamics. The proposed DCA model leverages discrete cosine transform (DCT) coefficients to incorporate frequency-domain information, capturing a broader spectrum of temporal features. Two selection criteria are employed to identify the most informative DCT coefficients for constructing the attention map. The first criterion utilizes the lowest frequency coefficients, whereas the second criterion selects the coefficients exhibiting the highest energy to construct the attention map. Comprehensive experiments on twelve diverse TSC datasets demonstrate that DCA consistently outperforms state-of-the-art attention mechanisms, achieving an average improvement of 2.2% in classification accuracy.</p><h2 dir="ltr">Other Information</h2><p dir="ltr">Published in: IEEE Open Journal of the Computer Society<br>License: <a href="https://creativecommons.org/licenses/by/4.0/deed.en" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/ojcs.2025.3586682" target="_blank">https://dx.doi.org/10.1109/ojcs.2025.3586682</a></p>2025-07-25T12:00:00ZTextJournal contributioninfo:eu-repo/semantics/publishedVersiontextcontribution to journal10.1109/ojcs.2025.3586682https://figshare.com/articles/journal_contribution/DCT-Based_Channel_Attention_for_Multivariate_Time_Series_Classification/30860261CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/308602612025-07-25T12:00:00Z
spellingShingle DCT-Based Channel Attention for Multivariate Time Series Classification
Amine Haboub (22392097)
Information and computing sciences
Artificial intelligence
Machine learning
Channel attention
convolutional neural networks
deep learning
discrete cosine transform
time series classification
Computational modeling
Feature extraction
Vectors
Training
Standards
Adaptation models
Accuracy
status_str publishedVersion
title DCT-Based Channel Attention for Multivariate Time Series Classification
title_full DCT-Based Channel Attention for Multivariate Time Series Classification
title_fullStr DCT-Based Channel Attention for Multivariate Time Series Classification
title_full_unstemmed DCT-Based Channel Attention for Multivariate Time Series Classification
title_short DCT-Based Channel Attention for Multivariate Time Series Classification
title_sort DCT-Based Channel Attention for Multivariate Time Series Classification
topic Information and computing sciences
Artificial intelligence
Machine learning
Channel attention
convolutional neural networks
deep learning
discrete cosine transform
time series classification
Computational modeling
Feature extraction
Vectors
Training
Standards
Adaptation models
Accuracy