{"paper_id":"2408.13214","title":"EUR-USD Exchange Rate Forecasting Based on Information Fusion with Large Language Models and Deep Learning Methods","authors":["Hongcheng Ding","Xuanze Zhao","Ruiting Deng","Shamsul Nahar Abdullah","Deshinta Arrova Dewi"],"paper_date":"2024-08-23","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2408.13214v2","page_url":"https://thequant.space/flowcharts/eur-usd-exchange-rate-forecasting-based-on-information-fusio/","scores":{"empirical_rigor":7.5,"math_complexity":4,"quadrant":"Street Traders"},"proposes":{"status":"available","statement":"The paper proposes the IUS framework that integrates unstructured textual and structured quantitative data, and an Optuna‑optimized Bi‑LSTM model for EUR/USD exchange rate forecasting.","quote":"We propose a novel IUS framework that integrates unstructured and structured data to predict exchange rates, setting a new benchmark in the industry.","summary":"Accurate forecasting of the EUR/USD exchange rate is crucial for investors, businesses, and policymakers. This paper proposes a novel framework, IUS, that integrates unstructured textual data from news and analysis with structured data on exchange rates and financial indicators to enhance exchange…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Missing quantitative values are linearly interpolated between known points.","type":"data","stated":"implied","quote":"We use linear interpolation to fill in these gaps: vi = va + (vb −va)(ti −ta) / (tb −ta)"}]},"datasets":{"status":"available","items":[{"name":"Textual dataset from investing.com and forexempire.com","provider":"investing.com, forexempire.com","kind":"text-news","access":"public","period":"2016-02-06 to 2024-01-19","universe":"35,427 texts (filtered to 20,329 relevant texts)","quote":"For the textual dataset, we collect data from investing.com and forexempire.com, covering the period from February 6, 2016, to January 19, 2024. The dataset includes all accessible data on these platforms within the period, totaling 35,427 texts"},{"name":"Quantitative financial data (exchange rates and market indicators)","provider":null,"kind":"market-prices","access":"unknown","period":null,"universe":null,"quote":"All collected raw financial data are fed into an indicator generator, which aligns the data and fills in missing values, producing the final quantitative features."}],"known":[{"id":"twitter","name":"Twitter / X","provider":"X Corp.","kind":"text-news","access":"public","quote":"This method provides deeper insights into market sentiment and enhances the predictive performance of their model, although it predominantly relies on Twitter, suggesting the potential benefits of incorporating more diverse data sources.","mentions":6,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"The Optuna‑Bi‑LSTM model reduces MAE by 10.69% and RMSE by 9.56% compared with the best baseline model.","kind":"performance","quote":"It can be observed that our proposed method consistently outperforms other models in terms of both MAE and RMSE metrics, achieving an improvement of at least 10.69% in MAE and 9.56% in RMSE compared with the best other model."},{"finding":"Combining unstructured textual and structured quantitative data yields higher accuracy than using structured data alone.","kind":"predictive-accuracy","quote":"In terms of data fusion, by combining unstructured and structured data, the model is able to enhance prediction accuracy beyond what is possible with structured data alone."},{"finding":"Using the top 12 important quantitative features together with 31 textual features is more effective than using all textual features.","kind":"predictive-accuracy","quote":"Furthermore, using the top 12 important features selected by the RFE method combined with 31 textual features proves to be more effective compared to analyzing all textual features."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":25,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T18:03:25Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}