EUR-USD Exchange Rate Forecasting Based on Information Fusion with Large Language Models and Deep Learning Methods

📅 2024-08-23
🏛️ arXiv.org
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
This paper addresses the limited accuracy of short-term EUR/USD exchange rate forecasting by proposing the IUS fusion framework. First, a large language model (LLM) performs dual-task understanding of financial news—simultaneously identifying sentiment polarity and classifying market trend direction. Second, a causality-driven feature generator aligns textual semantics with 12 key quantitative indicators (e.g., exchange rates, interest rates, volatility) via causal modeling and cross-modal integration. Finally, an Optuna-optimized bidirectional LSTM delivers end-to-end prediction. Key contributions include: (i) the first causal fusion mechanism bridging textual semantics and quantitative features; (ii) an LLM-powered dual-task paradigm for financial news understanding; and (iii) empirical validation of substantial performance gains from multi-source heterogeneous data integration. Experiments demonstrate that the IUS framework reduces MAE and RMSE by 10.69% and 9.56%, respectively, over the strongest baseline, confirming its superiority over purely structured modeling approaches.

Technology Category

Application Category

📝 Abstract
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 rate prediction. The IUS framework employs large language models for sentiment polarity scoring and exchange rate movement classification of texts. These textual features are combined with quantitative features and input into a Causality-Driven Feature Generator. An Optuna-optimized Bi-LSTM model is then used to forecast the EUR/USD exchange rate. Experiments demonstrate that the proposed method outperforms benchmark models, reducing MAE by 10.69% and RMSE by 9.56% compared to the best performing baseline. Results also show the benefits of data fusion, with the combination of unstructured and structured data yielding higher accuracy than structured data alone. Furthermore, feature selection using the top 12 important quantitative features combined with the textual features proves most effective. The proposed IUS framework and Optuna-Bi-LSTM model provide a powerful new approach for exchange rate forecasting through multi-source data integration.
Problem

Research questions and friction points this paper is trying to address.

Forecasting EUR/USD exchange rate using multi-source data fusion
Integrating unstructured text and structured financial data for prediction
Improving accuracy with LLM-based sentiment analysis and deep learning
Innovation

Methods, ideas, or system contributions that make the work stand out.

Large language models analyze text sentiment
Optuna-optimized Bi-LSTM for exchange rate prediction
Causality-Driven Feature Generator combines data types
🔎 Similar Papers
H
Hongcheng Ding
INTI International University, Malaysia
X
Xuanze Zhao
INTI International University, Malaysia
Z
Zixiao Jiang
INTI International University, Malaysia
S
Shamsul Nahar Abdullah
INTI International University, Malaysia
D
Deshinta Arrova Dewi
INTI International University, Malaysia