AMDCnet: attention-gate-based multi-scale decomposition and collaboration network for long-term time series forecasting

IntroductionTime series analysis plays a critical role in various applications, including sensor data monitoring, weather forecasting, economic predictions, and network traffic management. While traditional methods primarily focus on modeling time series data at a single temporal scale and achieve n...

Full description

Saved in:
Bibliographic Details
Main Authors: Shikang Hou, Song Sun, Tao Yin, Zhibin Zhang, Meng Yan
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-05-01
Series:Frontiers in Artificial Intelligence
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/frai.2025.1607232/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850122448774627328
author Shikang Hou
Song Sun
Tao Yin
Zhibin Zhang
Meng Yan
author_facet Shikang Hou
Song Sun
Tao Yin
Zhibin Zhang
Meng Yan
author_sort Shikang Hou
collection DOAJ
description IntroductionTime series analysis plays a critical role in various applications, including sensor data monitoring, weather forecasting, economic predictions, and network traffic management. While traditional methods primarily focus on modeling time series data at a single temporal scale and achieve notable results, they often overlook dependencies across multiple scales. Furthermore, the intricate structure of multi-scale time series complicates the effective extraction of features at different temporal resolutions.MethodTo address these limitations, we propose AMDCnet, a multi-scale-based time series decomposition and collaboration network designed to enhance the model's capacity for decomposing and integrating data across varying time scales. Specifically, AMDCnet transforms the original time series into multiple temporal resolutions and conducts multi-scale feature decomposition while preserving the overall temporal dynamics. By extracting features from downsampled sequences and integrating multi-resolution features through attention-gated co-training mechanisms, AMDCnet enables efficient modeling of complex time series data.ResultsAMDCnet achieving 44 best results and 10 second-best results out of 64 cases. Experimental results on 8 benchmark datasets demonstrate that AMDCnet achieves state-of-the-art performance in time series forecasting.DiscussionOur research provides a robust baseline for the application of artificial intelligence in multivariate time series forecasting. This work leverages multi-scale time series decomposition and gated units for feature fusion, effectively capturing dependencies across different temporal scales. Future studies may further optimize the scale decomposition and fusion modules. Such efforts could enhance the representation of multi-scale information and help address key challenges in multivariate time series prediction.
format Article
id doaj-art-1d377f8352f241ae953c3fb8d699e58c
institution OA Journals
issn 2624-8212
language English
publishDate 2025-05-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Artificial Intelligence
spelling doaj-art-1d377f8352f241ae953c3fb8d699e58c2025-08-20T02:34:50ZengFrontiers Media S.A.Frontiers in Artificial Intelligence2624-82122025-05-01810.3389/frai.2025.16072321607232AMDCnet: attention-gate-based multi-scale decomposition and collaboration network for long-term time series forecastingShikang Hou0Song Sun1Tao Yin2Zhibin Zhang3Meng Yan4School of Big Data and Software Engineering, Chongqing University, Chongqing, ChinaSchool of Computer and Information Science, Chongqing Normal University, Chongqing, ChinaSchool of Big Data and Software Engineering, Chongqing University, Chongqing, ChinaSchool of Big Data and Software Engineering, Chongqing University, Chongqing, ChinaSchool of Big Data and Software Engineering, Chongqing University, Chongqing, ChinaIntroductionTime series analysis plays a critical role in various applications, including sensor data monitoring, weather forecasting, economic predictions, and network traffic management. While traditional methods primarily focus on modeling time series data at a single temporal scale and achieve notable results, they often overlook dependencies across multiple scales. Furthermore, the intricate structure of multi-scale time series complicates the effective extraction of features at different temporal resolutions.MethodTo address these limitations, we propose AMDCnet, a multi-scale-based time series decomposition and collaboration network designed to enhance the model's capacity for decomposing and integrating data across varying time scales. Specifically, AMDCnet transforms the original time series into multiple temporal resolutions and conducts multi-scale feature decomposition while preserving the overall temporal dynamics. By extracting features from downsampled sequences and integrating multi-resolution features through attention-gated co-training mechanisms, AMDCnet enables efficient modeling of complex time series data.ResultsAMDCnet achieving 44 best results and 10 second-best results out of 64 cases. Experimental results on 8 benchmark datasets demonstrate that AMDCnet achieves state-of-the-art performance in time series forecasting.DiscussionOur research provides a robust baseline for the application of artificial intelligence in multivariate time series forecasting. This work leverages multi-scale time series decomposition and gated units for feature fusion, effectively capturing dependencies across different temporal scales. Future studies may further optimize the scale decomposition and fusion modules. Such efforts could enhance the representation of multi-scale information and help address key challenges in multivariate time series prediction.https://www.frontiersin.org/articles/10.3389/frai.2025.1607232/fulllong-term time seriesforecastingmulti-scale decompositionfeature fusionattention-gate
spellingShingle Shikang Hou
Song Sun
Tao Yin
Zhibin Zhang
Meng Yan
AMDCnet: attention-gate-based multi-scale decomposition and collaboration network for long-term time series forecasting
Frontiers in Artificial Intelligence
long-term time series
forecasting
multi-scale decomposition
feature fusion
attention-gate
title AMDCnet: attention-gate-based multi-scale decomposition and collaboration network for long-term time series forecasting
title_full AMDCnet: attention-gate-based multi-scale decomposition and collaboration network for long-term time series forecasting
title_fullStr AMDCnet: attention-gate-based multi-scale decomposition and collaboration network for long-term time series forecasting
title_full_unstemmed AMDCnet: attention-gate-based multi-scale decomposition and collaboration network for long-term time series forecasting
title_short AMDCnet: attention-gate-based multi-scale decomposition and collaboration network for long-term time series forecasting
title_sort amdcnet attention gate based multi scale decomposition and collaboration network for long term time series forecasting
topic long-term time series
forecasting
multi-scale decomposition
feature fusion
attention-gate
url https://www.frontiersin.org/articles/10.3389/frai.2025.1607232/full
work_keys_str_mv AT shikanghou amdcnetattentiongatebasedmultiscaledecompositionandcollaborationnetworkforlongtermtimeseriesforecasting
AT songsun amdcnetattentiongatebasedmultiscaledecompositionandcollaborationnetworkforlongtermtimeseriesforecasting
AT taoyin amdcnetattentiongatebasedmultiscaledecompositionandcollaborationnetworkforlongtermtimeseriesforecasting
AT zhibinzhang amdcnetattentiongatebasedmultiscaledecompositionandcollaborationnetworkforlongtermtimeseriesforecasting
AT mengyan amdcnetattentiongatebasedmultiscaledecompositionandcollaborationnetworkforlongtermtimeseriesforecasting