
Using Textual and Economic Features to Predict the RMB Exchange Rate
Author(s) -
Yi-Chen Chung,
Hsien-Ming Chou,
Chih-Neng Hung,
Chihli Hung
Publication year - 2021
Publication title -
advances in management and applied economics
Language(s) - English
Resource type - Journals
ISSN - 1792-7544
DOI - 10.47260/amae/1168
Subject(s) - renminbi , exchange rate , feature selection , support vector machine , computer science , econometrics , us dollar , ensemble forecasting , model selection , ensemble learning , artificial intelligence , regression , feature (linguistics) , regression analysis , selection (genetic algorithm) , machine learning , economics , statistics , mathematics , linguistics , philosophy , macroeconomics
This research proposes an integrated framework for the use of textual and economicfeatures to predict the exchange rate of the TWD (Taiwan dollar) against the RMB(Chinese Renminbi). The exchange rate is affected by the current economicsituation and expectations for the future economic climate. Exchange rateforecasting studies focus mainly on overall economic indices and the actualexchange rate, but overlook the influence of news. This research considers bothtextual and economic factors and builds three basic prediction models, i.e. multiplelinear regression (MLR), support vector regression (SVR), and Gaussian processregression (GPR) for the prediction of the RMB exchange rate. In addition to thethree basic prediction models, this research uses ensemble learning and featureselection techniques to improve prediction performance. Our experimentsdemonstrate that textual features also play an important role in predicting the RMBexchange rate. The SVR model is shown to outperform the other models and theMLR model is shown to perform worst. The ensemble of three basic modelsperforms better than its individual counterparts. Finally, the models which usefeature selection techniques demonstrate improved results in general, and differentfeature selection techniques are shown to be more suitable for different predictionmodels.JEL classification numbers: D80, F31, F47.Keywords: Exchange rate prediction, Text mining, Ensemble learning, Time seriesforecasting.