{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,3]],"date-time":"2026-03-03T10:00:17Z","timestamp":1772532017667,"version":"3.50.1"},"reference-count":43,"publisher":"PeerJ","license":[{"start":{"date-parts":[[2026,3,3]],"date-time":"2026-03-03T00:00:00Z","timestamp":1772496000000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Deanship of Scientific Research (DSR) at King Abdulaziz University, Jeddah, Saudi Arabia","award":["1285-611-2025"],"award-info":[{"award-number":["1285-611-2025"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>Long-term time series forecasting is essential for predicting future values based on historical patterns over extended periods. However, it poses significant challenges in effectively capturing both global trends and local variations. This research introduces FusionNet-FR, a novel DualNet Transformer architecture based on ensemble deep learning, which integrates fast-learning with slow-learning components to improve forecast accuracy. The fast learners are designed to extract high-frequency components and short-term variations, enabling the model to adapt to rapid changes in data patterns. In contrast, the slow learner is responsible for modeling long-term dependencies by leveraging global temporal structures through an inverted self-attention mechanism. In addition, incorporating frequency domain representation and interseries dependency modeling improves the robustness of the model against noisy data. To further strengthen learning, a controlled reconstruction loss strategy is integrated into the slow learner, facilitating self-supervised learning mechanisms. We evaluate FusionNet-FR on real-world multivariate and univariate datasets\u2014Exchange, ETT, Weather, and ILI\u2014using mean squared error (MSE) and mean absolute error (MAE) as performance metrics. FusionNet-FR achieves consistent improvements across both settings, with average percentage gains of 38.08% in MSE and 29.06% in MAE over baseline models on multivariate datasets, and 51.58% in MSE and 36.53% in MAE on univariate datasets. These results demonstrate the model\u2019s effectiveness in addressing the inherent challenges of long-term time series forecasting, delivering superior accuracy compared to existing approaches.<\/jats:p>","DOI":"10.7717\/peerj-cs.3620","type":"journal-article","created":{"date-parts":[[2026,3,3]],"date-time":"2026-03-03T08:24:29Z","timestamp":1772526269000},"page":"e3620","source":"Crossref","is-referenced-by-count":0,"title":["FusionNet-FR: a DualNet transformer architecture using frequency modeling and relational learning for enhanced long-term time series forecasting"],"prefix":"10.7717","volume":"12","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8762-3141","authenticated-orcid":true,"given":"Ary Mazharuddin","family":"Shiddiqi","sequence":"first","affiliation":[{"name":"Department of Informatics, Institut Teknologi Sepuluh Nopember, Surabaya, East Java, Indonesia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Andika Laksana","family":"Putra","sequence":"additional","affiliation":[{"name":"Department of Informatics, Institut Teknologi Sepuluh Nopember, Surabaya, East Java, Indonesia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ratih Nur Esti","family":"Anggraini","sequence":"additional","affiliation":[{"name":"Department of Informatics, Institut Teknologi Sepuluh Nopember, Surabaya, East Java, Indonesia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Farrela Ranku","family":"Mahhisa","sequence":"additional","affiliation":[{"name":"Department of Informatics, Institut Teknologi Sepuluh Nopember, Surabaya, East Java, Indonesia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Arif","family":"Bramantoro","sequence":"additional","affiliation":[{"name":"School of Computing and Informatics, Universiti Teknologi Brunei, Bandar Seri Begawan, Brunei Darussalam, Brunei"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3829-5970","authenticated-orcid":true,"given":"Fayruz","family":"Rahma","sequence":"additional","affiliation":[{"name":"Department of Informatics, Universitas Islam Indonesia, Sleman, Daerah Istimewa Yogyakarta, Indonesia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8081-6530","authenticated-orcid":true,"given":"Ahmad A.","family":"Alzahrani","sequence":"additional","affiliation":[{"name":"Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah, Makkah Province, Saudi Arabia"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"4443","published-online":{"date-parts":[[2026,3,3]]},"reference":[{"issue":"1","key":"10.7717\/peerj-cs.3620\/ref-1","doi-asserted-by":"publisher","first-page":"10","DOI":"10.5334\/dsj-2024-010","article-title":"KBJNet: kinematic bi-joint temporal convolutional network attention for anomaly detection in multivariate time series data","volume":"23","author":"Abdan Mulia","year":"2024","journal-title":"Data Science Journal"},{"issue":"6","key":"10.7717\/peerj-cs.3620\/ref-2","doi-asserted-by":"publisher","first-page":"6989","DOI":"10.1609\/aaai.v37i6.25854","article-title":"NHITS: neural hierarchical interpolation for time series forecasting","volume":"37","author":"Challu","year":"2023","journal-title":"2023 Proceedings of the 37th AAAI Conference on Artificial Intelligence, AAAI"},{"key":"10.7717\/peerj-cs.3620\/ref-3","first-page":"146","article-title":"Learning to rotate: quaternion transformer for complicated periodical time series forecasting","author":"Chen","year":"2022"},{"key":"10.7717\/peerj-cs.3620\/ref-4","first-page":"1437","article-title":"Formertime: hierarchical multi-scale representations for multivariate time series classification","author":"Cheng","year":"2023"},{"key":"10.7717\/peerj-cs.3620\/ref-5","first-page":"212","article-title":"TarNet: task-aware reconstruction for time-series transformer","author":"Chowdhury","year":"2022"},{"issue":"3","key":"10.7717\/peerj-cs.3620\/ref-6","doi-asserted-by":"publisher","first-page":"269","DOI":"10.1016\/j.neucom.2019.12.118","article-title":"Multivariate time series forecasting via attention-based encoder\u2013decoder framework","volume":"388","author":"Du","year":"2020","journal-title":"Neurocomputing"},{"key":"10.7717\/peerj-cs.3620\/ref-7","first-page":"459","article-title":"TSMixer: lightweight MLP-mixer model for multivariate time series forecasting","author":"Ekambaram","year":"2023"},{"key":"10.7717\/peerj-cs.3620\/ref-8","first-page":"544","article-title":"Sparse binary transformers for multivariate time series modeling","author":"Gorbett","year":"2023"},{"key":"10.7717\/peerj-cs.3620\/ref-9","article-title":"Reversible instance normalization for accurate time-series forecasting against distribution shift","author":"Kim","year":"2022"},{"key":"10.7717\/peerj-cs.3620\/ref-10","article-title":"Reformer: the efficient transformer","author":"Kitaev","year":"2020"},{"key":"10.7717\/peerj-cs.3620\/ref-11","doi-asserted-by":"publisher","first-page":"95","DOI":"10.1145\/3209978.3210006","article-title":"Modeling long- and short-term temporal patterns with deep neural networks","volume-title":"Proceedings of the 41st International ACM SIGIR Conference on Research & Development in Information Retrieval (SIGIR 2018),","author":"Lai","year":"2018"},{"key":"10.7717\/peerj-cs.3620\/ref-12","doi-asserted-by":"publisher","first-page":"5243","DOI":"10.5555\/3454287.3454758","article-title":"Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting","volume-title":"Proceedings of the 33rd International Conference on Neural Information Processing Systems","author":"Li","year":"2019a"},{"key":"10.7717\/peerj-cs.3620\/ref-13","volume-title":"Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting","author":"Li","year":"2019b"},{"key":"10.7717\/peerj-cs.3620\/ref-14","first-page":"2169","article-title":"Smartformer: semi-autoregressive transformer with efficient integrated window attention for long time series forecasting","author":"Li","year":"2023"},{"key":"10.7717\/peerj-cs.3620\/ref-15","doi-asserted-by":"publisher","DOI":"10.1109\/IJCNN64981.2025.11229402","article-title":"MTS-mixers: multivariate time series forecasting via factorized temporal and channel mixing","volume-title":"2025 International Joint Conference on Neural Networks (IJCNN), Rome, Italy","author":"Li","year":"2025"},{"issue":"1","key":"10.7717\/peerj-cs.3620\/ref-16","doi-asserted-by":"publisher","first-page":"12023","DOI":"10.1088\/1742-6596\/2209\/1\/012023","article-title":"Object detection based on YOLOv4-Tiny and improved bidirectional feature pyramid network","volume":"2209","author":"Liu","year":"2022a","journal-title":"Journal of Physics: Conference Series"},{"key":"10.7717\/peerj-cs.3620\/ref-17","article-title":"iTransformer: inverted transformers are effective for time series forecasting","author":"Liu","year":"2024"},{"key":"10.7717\/peerj-cs.3620\/ref-18","first-page":"9881","article-title":"Non-stationary transformers: exploring the stationarity in time series forecasting","volume-title":"Proceedings of the 36th International Conference on Neural Information Processing Systems (NIPS \u201922)","author":"Liu","year":"2022b"},{"key":"10.7717\/peerj-cs.3620\/ref-19","doi-asserted-by":"publisher","first-page":"91896","DOI":"10.1109\/access.2021.3091162","article-title":"Forecast methods for time series data: a survey","volume":"9","author":"Liu","year":"2021","journal-title":"IEEE Access"},{"issue":"8","key":"10.7717\/peerj-cs.3620\/ref-20","doi-asserted-by":"publisher","first-page":"4258","DOI":"10.1109\/tai.2024.3365775","article-title":"Dynamic long-term time-series forecasting via meta transformer networks","volume":"5","author":"Masum","year":"2024","journal-title":"IEEE Transactions on Artificial Intelligence"},{"key":"10.7717\/peerj-cs.3620\/ref-21","article-title":"A time series is worth 64 words: long-term forecasting with transformers","author":"Nie","year":"2023"},{"issue":"8","key":"10.7717\/peerj-cs.3620\/ref-22","doi-asserted-by":"publisher","first-page":"7733","DOI":"10.1007\/s10462-022-10319-w","article-title":"Time series forecasting using fuzzy cognitive maps: a survey","volume":"56","author":"Orang","year":"2023","journal-title":"Artificial Intelligence Review"},{"key":"10.7717\/peerj-cs.3620\/ref-23","article-title":"N-Beats: neural basis expansion analysis for interpretable time series forecasting","author":"Oreshkin","year":"2020"},{"key":"10.7717\/peerj-cs.3620\/ref-24","first-page":"16131","article-title":"DualNet: continual learning, fast and slow","volume-title":"Proceedings of the 35th International Conference on Neural Information Processing Systems (NIPS \u201921)","author":"Pham","year":"2021"},{"issue":"3","key":"10.7717\/peerj-cs.3620\/ref-25","doi-asserted-by":"publisher","first-page":"1181","DOI":"10.1016\/j.ijforecast.2019.07.001","article-title":"DeepAR: probabilistic forecasting with autoregressive recurrent networks","volume":"36","author":"Salinas","year":"2020","journal-title":"International Journal of Forecasting"},{"key":"10.7717\/peerj-cs.3620\/ref-26","first-page":"4837","article-title":"Think globally, act locally: a deep neural network approach to high-dimensional time series forecasting","volume-title":"Proceedings of the 33rd International Conference on Neural Information Processing Systems","author":"Sen","year":"2019"},{"key":"10.7717\/peerj-cs.3620\/ref-27","doi-asserted-by":"publisher","first-page":"24261","DOI":"10.48550\/arXiv.2105.01601","article-title":"MLP-mixer: an all-MLP architecture for vision","volume":"29","author":"Tolstikhin","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"10.7717\/peerj-cs.3620\/ref-28","article-title":"Attention is all you need","volume-title":"Advances in Neural Information Processing Systems","volume":"30","author":"Vaswani","year":"2017"},{"key":"10.7717\/peerj-cs.3620\/ref-29","article-title":"Multi-scale local and global context modeling for long-term series forecasting","author":"Wang","year":"2023"},{"key":"10.7717\/peerj-cs.3620\/ref-30","article-title":"Cost: contrastive learning of disentangled seasonal-trend representations for time series forecasting","author":"Woo","year":"2022"},{"key":"10.7717\/peerj-cs.3620\/ref-31","article-title":"TimesNet: temporal 2D-variation modeling for general time series analysis","author":"Wu","year":"2022"},{"key":"10.7717\/peerj-cs.3620\/ref-32","article-title":"Autoformer: decomposition transformers with auto-correlation for long-term series forecasting","author":"Wu","year":"2021"},{"key":"10.7717\/peerj-cs.3620\/ref-33","article-title":"Fits: modeling time series with 10k parameters","author":"Xu","year":"2023"},{"key":"10.7717\/peerj-cs.3620\/ref-34","first-page":"76656","article-title":"Frequency-domain MLPs are more effective learners in time series forecasting","volume-title":"Proceedings of the 37th International Conference on Neural Information Processing Systems (NIPS '23)","author":"Yi","year":"2024"},{"issue":"8","key":"10.7717\/peerj-cs.3620\/ref-35","doi-asserted-by":"publisher","first-page":"8980","DOI":"10.1609\/aaai.v36i8.20881","article-title":"TS2Vec: towards universal representation of time series","volume":"36","author":"Yue","year":"2022","journal-title":"2022 Proceedings of the 36th AAAI Conference on Artificial Intelligence, AAAI"},{"key":"10.7717\/peerj-cs.3620\/ref-36","first-page":"12310","article-title":"Barlow Twins: self-supervised learning via redundancy reduction","volume":"139","author":"Zbontar","year":"2021"},{"issue":"9","key":"10.7717\/peerj-cs.3620\/ref-37","doi-asserted-by":"publisher","first-page":"11121","DOI":"10.1609\/aaai.v37i9.26317","article-title":"Are transformers effective for time series forecasting?","volume":"37","author":"Zeng","year":"2022","journal-title":"Proceedings AAAI Conference Artificial Intelligence (AAAI)"},{"key":"10.7717\/peerj-cs.3620\/ref-38","first-page":"3586","article-title":"FRNet: frequency-based rotation network for long-term time series forecasting","author":"Zhang","year":"2024"},{"key":"10.7717\/peerj-cs.3620\/ref-39","article-title":"Crossformer: transformer utilizing cross-dimension dependency for multivariate time series forecasting","author":"Zhang","year":"2023"},{"key":"10.7717\/peerj-cs.3620\/ref-40","first-page":"12677","article-title":"FILM: frequency improved Legendre memory model for long-term time series forecasting","volume":"35","author":"Zhou","year":"2022a","journal-title":"Advances in Neural Information Processing Systems"},{"key":"10.7717\/peerj-cs.3620\/ref-41","first-page":"27268","article-title":"FEDformer: frequency enhanced decomposed transformer for long-term series forecasting","volume-title":"Proceedings 39th International Conference on Machine Learning (ICML), Baltimore, Maryland, USA, Jul. 17\u201323, 2022, Proceedings of Machine Learning Research","volume":"162","author":"Zhou","year":"2022b"},{"key":"10.7717\/peerj-cs.3620\/ref-42","article-title":"Informer: beyond efficient transformer for long sequence time-series forecasting","author":"Zhou","year":"2020"},{"issue":"12","key":"10.7717\/peerj-cs.3620\/ref-43","doi-asserted-by":"publisher","first-page":"11106","DOI":"10.1609\/aaai.v35i12.17325","article-title":"Informer: beyond efficient transformer for long sequence time-series forecasting","volume":"12B","author":"Zhou","year":"2021","journal-title":"35th AAAI Conference on Artificial Intelligence, AAAI 2021"}],"container-title":["PeerJ Computer Science"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/peerj.com\/articles\/cs-3620.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3620.xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3620.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3620.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,3]],"date-time":"2026-03-03T08:24:39Z","timestamp":1772526279000},"score":1,"resource":{"primary":{"URL":"https:\/\/peerj.com\/articles\/cs-3620"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,3]]},"references-count":43,"alternative-id":["10.7717\/peerj-cs.3620"],"URL":"https:\/\/doi.org\/10.7717\/peerj-cs.3620","archive":["CLOCKSS","LOCKSS","Portico"],"relation":{},"ISSN":["2376-5992"],"issn-type":[{"value":"2376-5992","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,3]]},"article-number":"e3620"}}