Adding double insurance to your investments: Evidence from the exchange rate market
Mingchen Li,Kun Yang,Zishu Cheng,Yunjie Wei,Shouyang Wang
DOI: https://doi.org/10.1016/j.aei.2024.102416
IF: 8.8
2024-04-01
Advanced Engineering Informatics
Abstract:The forecasting and trading of exchange rates is a prevalent and intricate undertaking within the realm of economics. While prior research has mostly concentrated on enhancing the accuracy of predictions through various methodologies, this study places greater emphasis on providing practical suggestions for effective trading. Hence, we propose a double insurance approach for exchange rate trading. Firstly, this study gathers time series data and textual information from various perspectives and markets. This enables us to generate point value forecasts by leveraging natural language processing and deep learning techniques. Simultaneously, this study employs an innovative interval forecasting method to track the upper and lower limits of daily exchange rate fluctuations. Finally, this study provides a novel trading method aimed at enhancing the possible financial gains associated with exchange rate fluctuations. The proposed methodology is validated through the utilization of exchange rates between the USD and four prominent currencies, namely CNY, JPY, EUR, and GBP. The findings indicate that the inclusion of economic variables and multi-platform sentiment indices enhances the precision of exchange rate prediction. Furthermore, the deep learning model (neural network model) exhibits superior performance compared to benchmark models (i.e. time series models and support vector regression). Furthermore, this study demonstrates that the implementation of the trading strategy with interval constraints significantly improves the profitability of exchange rate trading.
engineering, multidisciplinary,computer science, artificial intelligence