基于GS-ARIMA模型的原油股票成交量预测研究  

Research on Forecast of Crude Oil Stock Turnover Based on GS-ARIMA Model

作　　者：刘敏 侯俊华[1] Liu Min;Hou Junhua(College of Economic and Management East China Institute of Technology,Nanchang Jiangxi 330000)

机构地区：[1]东华理工大学经济与管理学院,江西南昌330000

出　　处：《对外经贸》2021年第12期38-40,共3页FOREIGN ECONOMIC RELATIONS & TRADE

摘　　要：针对ARIMA模型参数选取问题,提出一种基于网格搜索的ARIMA模型(GS-ARIMA)。选取2000-2019年原油股票成交量为研究样本,通过对数据误差分析、模型参数估计、模型检验等综合分析,建立了原油股票成交量的最优预测GS-ARIMA(6,1,5)模型。研究结果表明:GS-ARIMA(6,1,5)模型的相对误差大概率在2%~9%范围内波动,且模型拟合效果的决定系数R2为0.92,具有较高的预测精度。Aiming at the problem of ARIMA model parameter selection, a grid search-based ARIMA model(GSARIMA) is proposed. Selecting the crude oil stock trading volume from 2000 to 2019 as the research sample,through comprehensive analysis of data error analysis, model parameter estimation, model testing, etc., the optimal forecast of crude oil stock trading volume GS-ARIMA(6, 1, 5))Model. The research results show that the relative error of the GS-ARIMA(6,1,5) model fluctuates in the range of 2%-9%, and the coefficient of determination R2 of the model fitting effect is 0.92, which has high prediction accuracy.

关 键 词：检验 网格搜索 GS-ARIMA模型 预测 

分 类 号：F830.59[经济管理—金融学]