摘要
黄金期货价格时间序列具有复杂性、随机性和非线性的特点.引入“分而治之”思想,使用“分解-重构-集成”的方法对黄金期货的收盘价进行预测.首先,用分解方法对原序列进行分解;然后,用样本熵(SE)对分解序列进行重构;接着,用优化的极限学习机方法对重构序列进行预测;最后,对重构序列预测值进行集成,得到最终的预测值.实证表明,提出的模型优于基准模型,取得了较好的预测效果.
The time series of gold futures price has the characteristics of complexity,randomness and nonlinear.This paper introduces the idea of divide and conquer and uses the method of decomposition-reconstruction-integration to predict the closing price of gold futures.Firstly,the original sequence is decomposed by decomposition method.Secondly,sample entropy(SE)is used to reconstruct the decomposition sequence.Then,the optimized extreme learning machine method is used to predict the reconstructed sequence.Finally,the reconstructed sequence predicted values are integrated to get the final predicted values.The empirical results show that the model proposed in this paper is superior to the benchmark model and has a good prediction effect.
作者
何林芸
HE Lin-yun(School of Statistics,Lanzhou University of Finance and Economics,Lanzhou 730020,China)
出处
《兰州文理学院学报（自然科学版）》
2023年第1期35-39,共5页
Journal of Lanzhou University of Arts and Science(Natural Sciences)
关键词
黄金期货价格预测
分解集成
样本熵
神经网络
gold futures price forecast
decomposition and integration
sample entropy
neural network
