{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,19]],"date-time":"2026-05-19T18:09:03Z","timestamp":1779214143687,"version":"3.51.4"},"reference-count":44,"publisher":"Public Library of Science (PLoS)","issue":"5","license":[{"start":{"date-parts":[[2026,5,19]],"date-time":"2026-05-19T00:00:00Z","timestamp":1779148800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"\n Traffic forecasting plays a critical role in the field of urban planning. Yet, existing methods struggle with modeling complicated spatiotemporal dependencies and capturing long-term patterns due to their multiscale nature. In this paper, we present a novel framework named DG-LLM that leverages the advantages of decomposed temporal representations and adaptive spatial connectivity to model spatiotemporal dependencies. In this framework, traffic signals are decomposed into intrinsic modes, and dynamic graphs are learned for each mode to represent the spatial dependencies. These representations are then incorporated with pre-trained Large Language Models for effective long-range temporal dependency modeling. We conducted comprehensive experiments across six real-world traffic datasets spanning urban mobility systems and highway traffic networks and evaluated short- and long-term forecasting. Experimental results demonstrate that our framework provides significant improvements over state-of-the-art approaches, including benchmark graph- and LLM-based spatiotemporal forecasting models, even in long-term forecasting scenarios with severe temporal instability. Our model outperforms other methods by achieving\n \n \n \n \n \n 13<\/mml:mn>\n \u2212<\/mml:mo>\n 19<\/mml:mn>\n %<\/mml:mi>\n <\/mml:mrow>\n <\/mml:math>\n <\/jats:alternatives>\n <\/jats:inline-formula>\n improvements in MAE and\n \n \n \n \n \n 19<\/mml:mn>\n \u2212<\/mml:mo>\n 25<\/mml:mn>\n %<\/mml:mi>\n <\/mml:mrow>\n <\/mml:math>\n <\/jats:alternatives>\n <\/jats:inline-formula>\n in RMSE across all six benchmarks compared with baseline approaches. Additional analyses, including ablation studies, robustness to missing data, and zero-shot cross-dataset evaluation, further validate the effectiveness and generalization capability of the proposed framework.\n <\/jats:p>","DOI":"10.1371\/journal.pone.0349527","type":"journal-article","created":{"date-parts":[[2026,5,19]],"date-time":"2026-05-19T17:34:37Z","timestamp":1779212077000},"page":"e0349527","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["DG-LLM: Decomposition-based dynamic graph adaptation of large language models for spatiotemporal traffic forecasting"],"prefix":"10.1371","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0009-0005-4745-2762","authenticated-orcid":true,"given":"Sadia","family":"Tabassum","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5640-829X","authenticated-orcid":true,"given":"Naushin","family":"Nower","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"340","published-online":{"date-parts":[[2026,5,19]]},"reference":[{"key":"pone.0349527.ref001","unstructured":"Texas A&M Transportation Institute. 2025 Urban Mobility Report. 2025. https:\/\/mobility.tamu.edu\/umr\/"},{"key":"pone.0349527.ref002","doi-asserted-by":"crossref","first-page":"7070","DOI":"10.1007\/s10489-024-05562-3","article-title":"Probabilistic spatio-temporal graph convolutional network for traffic forecasting","volume":"54","author":"AA Karim","year":"2024","journal-title":"Appl Intell"},{"issue":"9","key":"pone.0349527.ref003","doi-asserted-by":"crossref","first-page":"3848","DOI":"10.1109\/TITS.2019.2935152","article-title":"T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction","volume":"21","author":"L Zhao","year":"2020","journal-title":"IEEE Trans Intell Transport Syst"},{"key":"pone.0349527.ref004","doi-asserted-by":"crossref","unstructured":"Rahman MdM, Nower N. Attention based Deep Hybrid Networks for Traffic Flow Prediction using Google Maps Data. In: Proceedings of the 2023 8th International Conference on Machine Learning Technologies, 2023. 74\u201381. https:\/\/doi.org\/10.1145\/3589883.3589894","DOI":"10.1145\/3589883.3589894"},{"key":"pone.0349527.ref005","doi-asserted-by":"crossref","unstructured":"Alghamdi T, Elgazzar K, Bayoumi M, Sharaf T, Shah S. Forecasting Traffic Congestion Using ARIMA Modeling. In: 2019 15th International Wireless Communications & Mobile Computing Conference (IWCMC), 2019. 1227\u201332. https:\/\/doi.org\/10.1109\/iwcmc.2019.8766698","DOI":"10.1109\/IWCMC.2019.8766698"},{"key":"pone.0349527.ref006","doi-asserted-by":"crossref","first-page":"582","DOI":"10.1016\/j.proeng.2017.04.417","article-title":"Traffic Flow Prediction using Kalman Filtering Technique","volume":"187","author":"SV Kumar","year":"2017","journal-title":"Procedia Engineering"},{"key":"pone.0349527.ref007","doi-asserted-by":"crossref","unstructured":"Liu C, Yang S, Xu Q, Li Z, Long C, Li Z, et al. Spatial-Temporal Large Language Model for Traffic Prediction. In: 2024 25th IEEE International Conference on Mobile Data Management (MDM), 2024. 31\u201340. https:\/\/doi.org\/10.1109\/mdm61037.2024.00025","DOI":"10.1109\/MDM61037.2024.00025"},{"key":"pone.0349527.ref008","unstructured":"Li Y, Yu R, Shahabi C, Liu Y. Diffusion convolutional recurrent neural network: data-driven traffic forecasting. arXiv preprint. 2017. https:\/\/doi.org\/10.48550\/arXiv.1707.01926"},{"key":"pone.0349527.ref009","doi-asserted-by":"crossref","unstructured":"Yu B, Yin H, Zhu Z. Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting. arXiv preprint. 2017. https:\/\/doi.org\/10.48550\/arXiv.1709.04875","DOI":"10.24963\/ijcai.2018\/505"},{"issue":"8","key":"pone.0349527.ref010","doi-asserted-by":"crossref","first-page":"4846","DOI":"10.1109\/TKDE.2025.3570705","article-title":"ST-LLM+: Graph Enhanced Spatio-Temporal Large Language Models for Traffic Prediction","volume":"37","author":"C Liu","year":"2025","journal-title":"IEEE Trans Knowl Data Eng"},{"issue":"3","key":"pone.0349527.ref011","doi-asserted-by":"crossref","first-page":"531","DOI":"10.1109\/TSP.2013.2288675","article-title":"Variational Mode Decomposition","volume":"62","author":"K Dragomiretskiy","year":"2014","journal-title":"IEEE Trans Signal Process"},{"key":"pone.0349527.ref012","doi-asserted-by":"crossref","unstructured":"Wu Z, Pan S, Long G, Jiang J, Zhang C. Graph wavenet for deep spatial-temporal graph modeling. arXiv preprint. 2019. https:\/\/arxiv.org\/abs\/1906.00121","DOI":"10.24963\/ijcai.2019\/264"},{"key":"pone.0349527.ref013","doi-asserted-by":"crossref","unstructured":"Wu Z, Pan S, Long G, Jiang J, Chang X, Zhang C. Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks. In: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2020. 753\u201363. https:\/\/doi.org\/10.1145\/3394486.3403118","DOI":"10.1145\/3394486.3403118"},{"key":"pone.0349527.ref014","first-page":"17804","article-title":"Adaptive graph convolutional recurrent network for traffic forecasting","volume":"33","author":"L Bai","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"issue":"01","key":"pone.0349527.ref015","doi-asserted-by":"crossref","first-page":"1234","DOI":"10.1609\/aaai.v34i01.5477","article-title":"GMAN: A Graph Multi-Attention Network for Traffic Prediction","volume":"34","author":"C Zheng","year":"2020","journal-title":"AAAI"},{"issue":"11","key":"pone.0349527.ref016","doi-asserted-by":"crossref","first-page":"5415","DOI":"10.1109\/TKDE.2021.3056502","article-title":"Learning Dynamics and Heterogeneity of Spatial-Temporal Graph Data for Traffic Forecasting","volume":"34","author":"S Guo","year":"2022","journal-title":"IEEE Trans Knowl Data Eng"},{"issue":"7","key":"pone.0349527.ref017","doi-asserted-by":"crossref","first-page":"485","DOI":"10.3390\/ijgi10070485","article-title":"A3T-GCN: Attention Temporal Graph Convolutional Network for Traffic Forecasting","volume":"10","author":"J Bai","year":"2021","journal-title":"IJGI"},{"key":"pone.0349527.ref018","doi-asserted-by":"crossref","unstructured":"Karim AA, Nower N. Robust Traffic Prediction Using Probabilistic Spatio-temporal Graph Convolutional Network. In: International Conference on Engineering Applications of Neural Networks. Springer Nature Switzerland; 2024. 259\u201373.","DOI":"10.1007\/978-3-031-62495-7_20"},{"issue":"1","key":"pone.0349527.ref019","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3532611","article-title":"Dynamic graph convolutional recurrent network for traffic prediction: benchmark and solution","volume":"17","author":"F Li","year":"2023","journal-title":"ACM Transactions on Knowledge Discovery from Data"},{"key":"pone.0349527.ref020","doi-asserted-by":"crossref","unstructured":"Han L, Du B, Sun L, Fu Y, Lv Y, Xiong H. Dynamic and Multi-faceted Spatio-temporal Deep Learning for Traffic Speed Forecasting. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, 2021. 547\u201355. https:\/\/doi.org\/10.1145\/3447548.3467275","DOI":"10.1145\/3447548.3467275"},{"issue":"6","key":"pone.0349527.ref021","doi-asserted-by":"crossref","first-page":"6367","DOI":"10.1609\/aaai.v36i6.20587","article-title":"Graph Neural Controlled Differential Equations for Traffic Forecasting","volume":"36","author":"J Choi","year":"2022","journal-title":"AAAI"},{"key":"pone.0349527.ref022","article-title":"Attention is all you need","volume":"30","author":"A Vaswani","year":"2017","journal-title":"Advances in neural information processing systems"},{"issue":"12","key":"pone.0349527.ref023","doi-asserted-by":"crossref","first-page":"11106","DOI":"10.1609\/aaai.v35i12.17325","article-title":"Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting","volume":"35","author":"H Zhou","year":"2021","journal-title":"AAAI"},{"key":"pone.0349527.ref024","first-page":"22419","article-title":"Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting","volume":"34","author":"H Wu","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"pone.0349527.ref025","doi-asserted-by":"crossref","unstructured":"Finzi M, Gruver N, Qiu S, Wilson A. Large Language Models Are Zero-Shot Time Series Forecasters. In: Advances in Neural Information Processing Systems 36, 2023. 19622\u201335. https:\/\/doi.org\/10.52202\/075280-0861","DOI":"10.52202\/075280-0861"},{"key":"pone.0349527.ref026","unstructured":"Jin M, Wang S, Ma L, Chu Z, Zhang JY, Shi X. Time-llm: Time series forecasting by reprogramming large language models. arXiv preprint. 2023. https:\/\/arxiv.org\/abs\/2310.01728"},{"issue":"11","key":"pone.0349527.ref027","doi-asserted-by":"crossref","first-page":"6851","DOI":"10.1109\/TKDE.2023.3342137","article-title":"PromptCast: A New Prompt-Based Learning Paradigm for Time Series Forecasting","volume":"36","author":"H Xue","year":"2024","journal-title":"IEEE Trans Knowl Data Eng"},{"issue":"3","key":"pone.0349527.ref028","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3719207","article-title":"LLM4TS: Aligning Pre-Trained LLMs as Data-Efficient Time-Series Forecasters","volume":"16","author":"C Chang","year":"2025","journal-title":"ACM Trans Intell Syst Technol"},{"key":"pone.0349527.ref029","doi-asserted-by":"crossref","unstructured":"Li Z, Xia L, Tang J, Xu Y, Shi L, Xia L, et al. UrbanGPT: Spatio-Temporal Large Language Models. In: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2024. 5351\u201362. https:\/\/doi.org\/10.1145\/3637528.3671578","DOI":"10.1145\/3637528.3671578"},{"key":"pone.0349527.ref030","doi-asserted-by":"crossref","unstructured":"Liu X, Hu J, Li Y, Diao S, Liang Y, Hooi B, et al. UniTime: A Language-Empowered Unified Model for Cross-Domain Time Series Forecasting. In: Proceedings of the ACM Web Conference 2024, 2024. 4095\u2013106. https:\/\/doi.org\/10.1145\/3589334.3645434","DOI":"10.1145\/3589334.3645434"},{"issue":"4","key":"pone.0349527.ref031","doi-asserted-by":"crossref","first-page":"1748","DOI":"10.1016\/j.ijforecast.2021.03.012","article-title":"Temporal Fusion Transformers for interpretable multi-horizon time series forecasting","volume":"37","author":"B Lim","year":"2021","journal-title":"International Journal of Forecasting"},{"key":"pone.0349527.ref032","doi-asserted-by":"crossref","unstructured":"Liu Y, Zhang Y, Zhang X, Tian L, Li Y, Luo J. UrbanMind: Urban Dynamics Prediction with Multifaceted Spatial-Temporal Large Language Models. In: Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2, 2025. 1951\u201362. https:\/\/doi.org\/10.1145\/3711896.3737177","DOI":"10.1145\/3711896.3737177"},{"key":"pone.0349527.ref033","unstructured":"Tabassum S, Ahmed S, Nower N. GAT-LLM: A graph attention-based framework with large language models for spatio-temporal traffic forecasting. In: 2025."},{"key":"pone.0349527.ref034","unstructured":"Yang N, Zhong H, Zhang H, Berry R. Vision-LLMs for Spatiotemporal Traffic Forecasting. In: 2025. https:\/\/doi.org\/arXiv:251011282"},{"issue":"8","key":"pone.0349527.ref035","doi-asserted-by":"crossref","first-page":"9035","DOI":"10.1609\/aaai.v38i8.28753","article-title":"ModWaveMLP: MLP-Based Mode Decomposition and Wavelet Denoising Model to Defeat Complex Structures in Traffic Forecasting","volume":"38","author":"K Sun","year":"2024","journal-title":"AAAI"},{"issue":"01","key":"pone.0349527.ref036","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1142\/S1793536909000047","article-title":"Ensemble Empirical Mode Decomposition: A Noise-assisted Data Analysis Method","volume":"01","author":"Z Wu","year":"2009","journal-title":"Adv Adapt Data Anal"},{"key":"pone.0349527.ref037","doi-asserted-by":"crossref","unstructured":"Torres ME, Colominas MA, Schlotthauer G, Flandrin P. A complete ensemble empirical mode decomposition with adaptive noise. In: 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2011. 4144\u20137. https:\/\/doi.org\/10.1109\/icassp.2011.5947265","DOI":"10.1109\/ICASSP.2011.5947265"},{"issue":"12","key":"pone.0349527.ref038","doi-asserted-by":"crossref","first-page":"7139","DOI":"10.3390\/app13127139","article-title":"A Long-Term Traffic Flow Prediction Model Based on Variational Mode Decomposition and Auto-Correlation Mechanism","volume":"13","author":"K Guo","year":"2023","journal-title":"Applied Sciences"},{"key":"pone.0349527.ref039","first-page":"100843","article-title":"An efficient and intelligent traffic flow prediction method based on LSTM and variational modal decomposition","volume":"28","author":"J Lu","year":"2023","journal-title":"Measurement: Sensors"},{"key":"pone.0349527.ref040","unstructured":"Veli\u010dkovi\u0107 P, Cucurull G, Casanova A, Romero A, Lio P, Bengio Y. Graph attention networks. arXiv preprint. 2017. https:\/\/arxiv.org\/abs\/1710.10903"},{"issue":"2","key":"pone.0349527.ref041","first-page":"3","article-title":"Lora: Low-rank adaptation of large language models","volume":"1","author":"EJ Hu","year":"2022","journal-title":"ICLR"},{"key":"pone.0349527.ref042","unstructured":"NYC Taxi and Limousine Commission. TLC Trip Record Data. https:\/\/www1.nyc.gov\/site\/tlc\/about\/tlc-trip-record-data.page 2025 December 26."},{"key":"pone.0349527.ref043","unstructured":"Citi Bike NYC. Citi Bike System Data. https:\/\/citibikenyc.com\/system-data 2024."},{"key":"pone.0349527.ref044","unstructured":"California Department of Transportation. Performance Measurement System (PeMS) Data Source. https:\/\/dot.ca.gov\/programs\/traffic-operations\/mpr\/pems-source 2025 December 26."}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0349527","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,19]],"date-time":"2026-05-19T17:34:49Z","timestamp":1779212089000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0349527"}},"subtitle":[],"editor":[{"given":"Shih-Lin","family":"Lin","sequence":"first","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2026,5,19]]},"references-count":44,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2026,5,19]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0349527","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,19]]}}}