{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:52:00Z","timestamp":1791525120262,"version":"4.3.4"},"reference-count":18,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T00:00:00Z","timestamp":1791504000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>Accurate financial time-series forecasting is essential for quantitative investment. However, existing methods often perform inadequately when handling the complex dynamics of financial data, including non-stationarity, volatility clustering, and market regime transitions. These characteristics can adversely affect both forecasting accuracy and the quality of investment decisions. To address these challenges, this study develops AHAT-Fin, an improved Transformer-based forecasting framework that integrates adaptive hybrid attention with market state perception. AHAT-Fin integrates dual-path adaptive hybrid attention, generalized autoregressive conditional heteroskedasticity (GARCH)-based dynamic volatility adjustment, and hierarchical market-state perception based on multi-window long short-term memory (LSTM) networks. The model is evaluated using daily trading data from seven representative U.S. stocks and selected constituents of the CSI 300 Index. The results show that AHAT-Fin generally outperforms the benchmark models, achieving an average reduction of 20.2% in mean squared error, an 8.4% improvement in directional accuracy, and a 63.4% increase in the Sharpe ratio. Module-level evaluation further confirms that all three components contribute to the overall predictive performance, with the market state perception module playing a particularly important role in improving risk-adjusted investment performance. Visual analyses of attention weights and market state identification further demonstrate the model\u2019s ability to extract multidimensional financial features and capture transitions across market states. These findings demonstrate the potential of combining volatility-aware modeling, adaptive attention, and market-state perception for financial forecasting and quantitative strategy design.<\/jats:p>","DOI":"10.3389\/frai.2026.1975632","type":"journal-article","created":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:33:24Z","timestamp":1791524004000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["AHAT-Fin: an adaptive hybrid attention and market-state-aware framework for financial time-series forecasting"],"prefix":"10.3389","volume":"9","author":[{"given":"Yingli","family":"Gong","sequence":"first","affiliation":[{"name":"School of Supply Chain Management, Ningbo Polytechnic University","place":["Ningbo, Zhejiang, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dongdong","family":"Lv","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence, Ningbo University of Technology","place":["Ningbo, Zhejiang, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhijian","family":"Duan","sequence":"additional","affiliation":[{"name":"School of Artificial Intelligence, Ningbo University of Technology","place":["Ningbo, Zhejiang, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,10,9]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"307","DOI":"10.1016\/0304-4076(86)90063-1","article-title":"Generalized autoregressive conditional heteroskedasticity","volume":"31","author":"Bollerslev","year":"1986","journal-title":"J. Econom."},{"key":"ref2","doi-asserted-by":"publisher","first-page":"211","DOI":"10.1111\/j.2517-6161.1964.tb00553.x","article-title":"An analysis of transformations","volume":"26","author":"Box","year":"1964","journal-title":"J. R. Stat. Soc. B"},{"key":"ref3","doi-asserted-by":"crossref","first-page":"1949","DOI":"10.1007\/s10614-023-10390-7","article-title":"Garchnet: value-at-risk forecasting with Garch models based on neural networks: M. Buczynski, M. Chlebus","volume":"63","author":"Buczynski","year":"2024","journal-title":"Comput. Econ."},{"key":"ref4","doi-asserted-by":"publisher","first-page":"137","DOI":"10.9781\/ijimai.2023.06.003","article-title":"Machine learning for financial prediction under regime change using technical analysis: a systematic review","volume":"9","author":"Cetrulo","year":"2024","journal-title":"Int. J. Interact. Multimed. Artif. Intell."},{"key":"ref5","doi-asserted-by":"publisher","first-page":"987","DOI":"10.2307\/1912773","article-title":"Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation","volume":"50","author":"Engle","year":"1982","journal-title":"Econometrica"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"19569","DOI":"10.1007\/s13369-025-10242-6","article-title":"Time series forecasting using transfer learning with an attention-revamped transformer: a case study of financial instruments","volume":"50","author":"Feng","year":"2025","journal-title":"Arab. J. Sci. Eng."},{"key":"ref7","doi-asserted-by":"publisher","first-page":"72","DOI":"10.1007\/s44443-025-00045-y","article-title":"A novel transformer-based dual attention architecture for the prediction of financial time series","volume":"37","author":"Hadizadeh","year":"2025","journal-title":"J. King Saud Univ. Comput. Inf. Sci."},{"key":"ref8","doi-asserted-by":"publisher","first-page":"357","DOI":"10.2307\/1912559","article-title":"A new approach to the economic analysis of nonstationary time series and the business cycle","volume":"57","author":"Hamilton","year":"1989","journal-title":"Econometrica"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"Hochreiter","year":"1997","journal-title":"Neural Comput."},{"key":"ref10","doi-asserted-by":"publisher","first-page":"7","DOI":"10.3390\/sci7010007","article-title":"LSTM\u2013transformer-based robust hybrid deep learning model for financial time series forecasting","volume":"7","author":"Kabir","year":"2025","journal-title":"Sci"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"313","DOI":"10.1007\/s10614-025-11042-8","article-title":"The sentiment augmented GARCH-LSTM hybrid model for value-at-risk forecasting","volume":"67","author":"L\u00e9ber","year":"2026","journal-title":"Comput. Econ."},{"key":"ref12","doi-asserted-by":"publisher","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":"Lim","year":"2021","journal-title":"Int. J. Forecast."},{"key":"ref13","doi-asserted-by":"publisher","first-page":"11116","DOI":"10.48550\/arXiv.2310.06625","article-title":"iTransformer: inverted transformers are effective for time series forecasting","author":"Liu","year":"2024","journal-title":"Int. Conf. Learn. Repr."},{"key":"ref14","doi-asserted-by":"publisher","first-page":"108223","DOI":"10.1016\/j.engappai.2024.108223","article-title":"Volatility forecasting and assessing risk of financial markets using multi-transformer neural network based architecture","volume":"133","author":"Mishra","year":"2024","journal-title":"Eng. Appl. Artif. Intell."},{"key":"ref15","doi-asserted-by":"crossref","first-page":"8553","DOI":"10.1007\/s00500-022-07161-1","article-title":"Deep learning for volatility forecasting in asset management: A. Petrozziello et al","volume":"26","author":"Petrozziello","year":"2022","journal-title":"Soft. Comput."},{"key":"ref16","doi-asserted-by":"publisher","first-page":"112600","DOI":"10.1016\/j.asoc.2024.112600","article-title":"Temporal fusion point-interval forecasting: a comprehensive approach for financial time series prediction","volume":"169","author":"Qi","year":"2025","journal-title":"Appl. Soft Comput."},{"key":"ref17","doi-asserted-by":"publisher","first-page":"723","DOI":"10.1007\/s12197-023-09629-8","article-title":"Do artificial neural networks provide improved volatility forecasts: evidence from Asian markets","volume":"47","author":"Sahiner","year":"2023","journal-title":"J. Econ. Finance"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"36","DOI":"10.1186\/s40854-023-00564-5","article-title":"Forecasting var and es by using deep quantile regression, Gans-based scenario generation, and heterogeneous market hypothesis","volume":"10","author":"Wang","year":"2024","journal-title":"Financ. Innov."}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1975632\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,9]],"date-time":"2026-10-09T05:33:25Z","timestamp":1791524005000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1975632\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,9]]},"references-count":18,"alternative-id":["10.3389\/frai.2026.1975632"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1975632","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,9]]},"article-number":"1975632"}}