RNN在线学习框架下CNN-LSTM模型对黄金期货价格的预测  

Prediction of Gold Future Prices Using CNN-LSTM Model under RNN Online Learning Framework

作　　者：石岩松 杨博 SHI Yansong;YANG Bo(School of Mathematics,Yunnan Normal University,Kunming 650500,China;Yunnan Key Laboratory of Modern Analytical Mathematics and Applications,Yunnan Normal University,Kunming 650500,China)

机构地区：[1]云南师范大学数学学院,云南昆明650500 [2]云南师范大学云南省现代分析数学及其应用重点实验室,云南昆明650500

出　　处：《现代信息科技》2024年第11期141-144,152,共5页Modern Information Technology

摘　　要：黄金是一种特殊的金融商品,具有避险功能。黄金期货价格受多方面因素的影响,一般认为黄金期货价格变化趋势呈现非线性非平稳的时间序列,传统的预测模型难以对其进行有效的预测。文章向传统在线学习算法中加入信息传递,提出基于RNN的在线学习算法ROA(RNN-based Online Algorithm);选用芝加哥商品交易所黄金期货价格数据进行实证分析,使用CNN-LSTM作为基础预测模型,以MAE、RMSE、R^(2)作为评价指标,结果表明在所有评价指标中ROA的预测性能均优于传统在线学习算法。Gold is a special financial commodity with a safe-haven function.The price of gold futures is affected by many factors and is generally regarded as a non-linear and non-stationary time series,which is difficult to be predicted by traditional forecasting models.We introduce information transmission into traditional online learning algorithms and proposes an online learning algorithm ROA(RNN based Online Algorithm)based on RNN(Recurrent Neural Network).Empirical analysis is conducted using the Chicago Mercantile Exchange gold futures price data,with CNN-LSTM(Convolutional Neural Networks-Long Short Term Memory)as the basic prediction model and MAE(Mean Absolute Error),RMSE(Root Mean Square Error),and R^(2) as evaluation indicators.The results show that the predictive performance of ROA is superior to traditional online learning algorithms in all evaluation indicators.

关 键 词：RNN 黄金期货价格 在线学习算法 

分 类 号：TP39[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]