We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
基于ARIMA时间序列模型的稀土氧化物价格预测研究.
- Authors
杨斌清; 张希琳
- Abstract
Owing to various impact factors with complicated interactions on the price of rare earth products,and the nonstationarity,nonlinear and noise characteristics of the monthly prices of rare earth products,the difficulty to forecast the price increases. In order to solve this problem,time series prediction method was adopted in this paper and a case study of the monthly prices of rare earths neodymium oxide and dysprosium oxide were analyzed.Through building nonstationarity ARIMA( 1,1,2) model to describe and forecast the dynamic fluctuation of price of rare earths neodymium oxide and dysprosium oxide,we obtained their prediction data from January 2006 to December 2015. Besides,the prediction data was compared with observation data by prediction fitting accuracy analysis.The result showed that the model had relatively high fitting accuracy,which is a good way to forecast prices of rare earth products. The results showed that the model had high fitting accuracy,which is suitable for short-term simulation of prices of rare earth neodymium oxide and dysprosium oxide product. Due to the similarity in the fluctuation period of the prices of major rare earth products in China,which means the close correlation between their markets,and the model can also be used to simulate the price of other rare earth oxides. The model can be applied to solve the real-world problem. For example,it can be used by rare earth industry management departments to compile price forecast report for the rare earth product regularly so that operators for rare earth production,consumers and governments are capable of acquiring trends and prices for rare earth products market and making prompt decisions.
- Subjects
NEODYMIUM compounds; DYSPROSIUM compounds; OXIDES; RARE earth oxides; TIME series analysis; PRICES
- Publication
Journal of the Chinese Society of Rare Earths, 2017, Vol 35, Issue 5, p680
- ISSN
1000-4343
- Publication type
Article
- DOI
10.11785/S1000-4343.20170515