Affiliation:
1. School of Engineering, Computer and Mathematical Sciences, Auckland University of Technology, Auckland 1010, New Zealand
Abstract
The currency exchange rate is a crucial link between all countries related to economic and trade activities. With increasing volatility, exchange rate fluctuations have become frequent under the combined effects of global economic uncertainty and political risks. Consequently, accurate exchange rate prediction is significant in managing financial risks and economic instability. In recent years, the Transformer models have attracted attention in the field of time series analysis. Transformer models, such as Informer and TFT (Temporal Fusion Transformer), have also been extensively studied. In this paper, we evaluate the performance of the Transformer, Informer, and TFT models based on four exchange rate datasets: NZD/USD, NZD/CNY, NZD/GBP, and NZD/AUD. The results indicate that the TFT model has achieved the highest accuracy in exchange rate prediction, with an R2 value of up to 0.94 and the lowest RMSE and MAE errors. However, the Informer model offers faster training and convergence speeds than the TFT and Transformer, making it more efficient. Furthermore, our experiments on the TFT model demonstrate that integrating the VIX index can enhance the accuracy of exchange rate predictions.
Reference26 articles.
1. Bi, Jiarui, Zhu, Zengliang, and Meng, Qinglong (, January September). Transformer in computer vision. Paper presented at 2021 IEEE International Conference on Computer Science, Electronic Information Engineering and Intelligent Control Technology (CEI), Fuzhou, China.
2. A temporal fusion transformer deep learning model for long-term streamflow forecasting: A case study in the funil reservoir;Fayer;Southeast Brazil. Knowledge-Based Engineering and Sciences,2023
3. Load forecasting of district heating system based on Informer;Gong;Energy,2022
4. A survey on vision transformer;Han;IEEE Transactions on Pattern Analysis and Machine Intelligence,2022
5. Hu, Zexin, Zhao, Yiqi, and Khushi, Matloob (2021). A survey of forex and stock price prediction using deep learning. Applied System Innovation, 4.