Exchange Rate Forecasting Using Machine learning: Explore Gains From External Information
Exchange Rate Forecasting Using Machine learning: Explore Gains From External Information
复制标题
DOI：
10.1145/3490700.3490704
复制
发表时间：
2021-09
期刊：
影响因子：
--
通讯作者：
Zhenlin Liang;Xiang Li
中科院分区：
文献类型：
--
作者：
Zhenlin Liang;Xiang Li
汇率市场是全球最大的金融市场，汇率的波动关系到外汇投资者的利益，对外贸公司的订单定价也至关重要。然而，仅凭历史数据进行预测并不足以支持决策，还需要外部信息。本文采用技术指标作为历史数据的特征，采用基础分析方法收集详细、全面的外部信息，采用集成特征分级器（Ensemble Feature Grader， EFG）过滤特征中的噪声。经过EFG方法的特征选择，可以提高模型的性能。与仅使用历史数据的模型相比，外部信息带来了显著的增益。
The exchange rate market is the world's largest financial market, and volatility in the exchange rate are relevant to forex investors' interests and crucial to the pricing of orders for foreign trade companies. However, forecasting with historical data alone is not enough to support decision-making and requires external information. In this paper, we use technical indicators as features of historical data, use fundamental analysis to collect detailed and comprehensive external information, and use Ensemble Feature Grader (EFG) to filter out the noise in the features. After EFG method feature selection, the model performance can be improved. The external information brings a significant gain compared to the model that uses only historical data.