基于长短期记忆网络的黄金价格预测  

Gold Price Prediction Based on Long Short-Term Memory Network

作　　者：闫铭 李东喜 

机构地区：[1]太原理工大学数学学院,山西 晋中 [2]太原理工大学大数据学院,山西 晋中

出　　处：《应用数学进展》2020年第6期871-880,共10页Advances in Applied Mathematics

摘　　要：精确的预测黄金价格,有助于投资者了解黄金市场的行情,并对他们做出正确的投资决策提供了科学有效的参考,因此,提高黄金价格的预测精度显得尤为重要。本文提出了一种基于长短期记忆网络(LSTM)模型的黄金价格预测方法。该方法结合黄金价格数据,利用长短期记忆网络生成训练模型,最终实现对黄金价格的预测。结果表明,本文的方法可行有效,较BP神经网络和SVR智能预测拥有更高的预测精度。Accurate prediction of gold price is helpful for investors to understand the gold market and provide scientific and effective reference for them to make correct investment decisions. Therefore, it is particularly important to improve the prediction accuracy of gold price. This paper presents a gold price forecasting method based on Long Short-term memory network (LSTM) model. This method combines the gold price data, uses LSTM model to generate training model, and finally realizes the prediction of gold price. The results show that this method is feasible and effective, and has higher prediction accuracy than BP neural network and SVR intelligent prediction.

关 键 词：黄金价格 长短期记忆网络(LSTM) 短期预测 长期预测 

分 类 号：F83[经济管理—金融学]