XTFT: A Next-Generation Temporal Fusion Transformer for Uncertainty-Rich Time Series Forecasting

XTFT: A Next-Generation Temporal Fusion Transformer for Uncertainty-Rich Time Series Forecasting | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 30 July 2025 V1 Latest version Share on XTFT: A Next-Generation Temporal Fusion Transformer for Uncertainty-Rich Time Series Forecasting Authors : Kouao Laurent Kouadio 0000-0001-7259-7254 [email protected] , Zhuo Liu , Rong Liu , Pierre Claver Bizimana , Gaofeng Yang , and Wenxiang Liu Authors Info & Affiliations https://doi.org/10.22541/au.175390529.91420978/v1 330 views 221 downloads Contents Abstract Supplementary Material Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Time-series forecasting struggles with multiscale dynamics, heterogeneous data, and uncertainty. We introduce the Extreme Temporal Fusion Transformer (XTFT) to address these limitations using multiscale LSTMs, learned normalization, multimodal embeddings, and enhanced attention. XTFT provides deterministic forecasts and reliable uncertainty estimates via its probabilistic framework. Evaluated on Nansha land subsidence prediction, XTFT significantly outperformed the Temporal Fusion Transformer (TFT), achieving MAE/RMSE of 11.36/14.46 mm (vs. TFT's 17.97/22.58 mm) and R2 of 0.888 (vs. 0.727). XTFT demonstrated robust probabilistic forecasts with high coverage rates (e.g., 88.58%). These results show XTFT effectively captures complex patterns and manages diverse data, setting new benchmarks for accuracy, interpretability, and efficiency in multihorizon time series forecasting. Note: This work is currently under review at the IEEE Transactions on Pattern Analysis and Machine Intelligence. Supplementary Material File (manuscript.pdf) Download 5.06 MB File (supplemental_material.pdf) Download 439.34 KB Information & Authors Information Version history V1 Version 1 30 July 2025 Copyright This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License Keywords hierarchical attention multi-horizon forecasting multi-modal embeddings probabilistic forecasting temporal dependencies xtft Authors Affiliations Kouao Laurent Kouadio 0000-0001-7259-7254 [email protected] School of Geosciences and Info-physics, Central South University, Changsha, Hunan, 410083, China View all articles by this author Zhuo Liu Department of Earth and Planetary Science, Stanford University, California 94305, USA View all articles by this author Rong Liu School of Geosciences and Info-physics, Central South University, Changsha, Hunan, 410083, China View all articles by this author Pierre Claver Bizimana School of Computer Science and Engineering, Central South University, Changsha, Hunan, 410083, China View all articles by this author Gaofeng Yang School of Geosciences and Info-physics, Central South University, Changsha, Hunan, 410083, China View all articles by this author Wenxiang Liu Guangdong Geological Bureau, Guangzhou, Guangdong, 510700, China View all articles by this author Funding Information National Science Foundation Metrics & Citations Metrics Article Usage 330 views 221 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Kouao Laurent Kouadio, Zhuo Liu, Rong Liu, et al. XTFT: A Next-Generation Temporal Fusion Transformer for Uncertainty-Rich Time Series Forecasting. Authorea . 30 July 2025. DOI: https://doi.org/10.22541/au.175390529.91420978/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); Cited by Kouao Laurent Kouadio, k-diagram: Rethinking Forecasting Uncertainty via Polar-based Visualization, Journal of Open Source Software, 10 , 116, (8661), (2025). https://doi.org/10.21105/joss.08661 Crossref Loading... View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.175390529.91420978/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9feb505bdae41640',t:'MTc3OTI3OTgyMA=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();