摘要
分析了全球铀资源情况,对国际天然铀市场历史波动和影响因素进行了梳理,认为国际天然铀市场经历了3个大的周期,每个周期的跨度都长达25—30年,可以利用国际原油、煤炭以及自身走势的惯性等因素对天然铀价格进行预测。在综合分析基础上,利用神经网络模型和自回归模型对国际天然铀年度和月度价格进行了预测研究,认为利用滞后一期的影响因素年度数据构建神经网络模型可捕捉到天然铀价格趋势性走势,选用一阶差分后的月度数据进行短期铀价预测可得到较好的效果。
This article analyzes the global uranium resources situation and sorts out the historical fluctuations and influencing factors of the international natural uranium market.This paper argues that the international natural uranium market has experienced three large cycles,each covering 25-30 years,and that natural uranium prices can be predicted using international crude oil,coal,the inentia of its own trend and some other factors.Based on comprehensive analysis,neural network models and ARIMA model were used to predict the annual and monthly prices of international natural uranium.This paper found that a neural network model using influencing factors’annual data with a one-period lag can capture the trend of uranium prices,while the firstorder differenced monthly data can also be used to predict short-term fluctuations with good results.
作者
沈达
张潇潭
车永飞
马娜
SHEN Da;ZHANG Xiaotan;CHE Yongfei;MA Na(China Nuclear Uranium Co.Ltd,Beijing 100013,China;University of Economics and Business,Beijing 100070,China;National Key Laboratory of Science and Technology on Remote Sensing Information and Imagery Analysis,Beijing Research Institute of Uranium Geology,Beijing,100029 China)
出处
《世界核地质科学》
CAS
2023年第2期272-280,共9页
World Nuclear Geoscience
基金
中国核工业地质局“柴达木盆地铀成矿信息高分卫星遥感综合识别与技术应用”项目(编号:遥D2305)资助。
