Forecasting US dollar exchange rate movement with computational models and human behavior
作者：
Highlights：
• Benefits are adding Behavioral Finance to the Machine Learning framework.
• We assume Calendar Effects induce deterministic patterns in financial time series.
• We used it to improve existing voting-based ensemble models with no retraining.
• Prediction: the Brazilian Real to the US Dollar daily exchange rate movement.
• The metric reached a value 24% higher than the original voting-based ensemble model.
摘要
•Benefits are adding Behavioral Finance to the Machine Learning framework.•We assume Calendar Effects induce deterministic patterns in financial time series.•We used it to improve existing voting-based ensemble models with no retraining.•Prediction: the Brazilian Real to the US Dollar daily exchange rate movement.•The metric reached a value 24% higher than the original voting-based ensemble model.
论文关键词：Exchange rate,Behavioral finance,Ensemble models,Machine learning
论文评审过程：Received 9 July 2020, Revised 4 January 2022, Accepted 7 January 2022, Available online 18 January 2022, Version of Record 29 January 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.116521
