基于ARIMA时间序列模型的稀土氧化物价格预测研究

发布时间：2018-04-29 07:13

本文选题：稀土价格 + 价格预测　；参考：《中国稀土学报》2017年05期

【摘要】：由于影响稀土产品价格的因素众多,各因素之间保持着错综复杂的联系,并且稀土产品月度价格数据具有高度的非平稳性、非线性和噪声的特性,增加了稀土产品价格预测难度。为此,运用时间序列预测法,以稀土Nd_2O_3,Dy_2O_3月度价格为例,建立非平稳时间序列ARIMA(1,1,2)模型,来描述并预测稀土产品价格的动态变化,得到了2006年1月~2015年12月的Nd_2O_3,Dy_2O_3价格预测值,并使用真实观测值与预测值进行预测拟合精度分析,结果表明,该模型拟合精度较高,适合于中短期模拟预测Nd_2O_3,Dy_2O_3产品价格。由于中国主要稀土产品价格的波动周期具有一定的相似性,表明各市场之间的关联比较密切,因此,该模型也可以用来模拟预测其他稀土氧化物的价格。该模型具有一定的实践运用价值,稀土行业管理部门可以运用该模型定期编制稀土产品价格预测报告,以便稀土产品生产经营者、消费者和各级政府随时掌握稀土产品市场价格情况和变动趋势,及时进行决策。
[Abstract]:Because of the large number of factors affecting the price of rare earth products, the various factors maintain a complex relationship, and the monthly price data of rare earth products have highly non-stationary, nonlinear and noise characteristics. It increases the difficulty of price prediction of rare earth products. Therefore, using the method of time series prediction and taking the monthly price of Nd2O3 / Dy _ 2O _ 3 as an example, a non-stationary time series (ARIMA) model is established to describe and predict the dynamic changes of the prices of rare earth products. The price prediction values of Nd2O3Dy2O3 are obtained from January 2006 to December 2015. The results show that the model has high fitting precision and is suitable for the short and medium term simulation to predict the price of NdSZ _ 2O _ 3 Dy _ 2O _ 3 products. Because the fluctuation cycle of the main rare earth products in China has certain similarity, which indicates that the relationship between the markets is close, the model can also be used to simulate and predict the prices of other rare earth oxides. The model has some practical application value. The management department of rare earth industry can use the model to compile the price forecast report of rare earth products regularly, so that the producers and operators of rare earth products can make use of the model. Consumers and governments at all levels keep track of market prices and trends of rare earth products and make timely decisions.
【作者单位】：江西理工大学应用科学学院;密歇根大学自然资源与环境学院;
【基金】：国家自然科学基金项目(71241022);国家自然科学基金项目(41462016) 江西省人文社科重点招标课题(jd1471)资助
【分类号】：F426;F764

【相似文献】