基于时间序列的上证综合指数短期预测分析  

Short-term Prediction Analysis of Shanghai Composite Index Based on Time Series

作　　者：刘亭[1] 赵月旭[1] 

机构地区：[1]杭州电子科技大学理学院,浙江杭州310018

出　　处：《杭州电子科技大学学报（自然科学版）》2017年第4期71-74,共4页Journal of Hangzhou Dianzi University：Natural Sciences

基　　金：国家自然科学基金资助项目(61273093;61473107;U1509205);浙江省自然科学基金资助项目(LR16F030003)

摘　　要：对上证日收盘指数2000-12-20至2016-06-20的数据进行建模及预测分析.首先利用相关数据建立ARIMA模型,发现模型的残差存在条件异方差性以及非正态性,于是对残差建立GARCH模型,并对残差的分布类型分别做正态分布、广义误差分布与t分布假设.通过预测精度对比发现,残差服从t分布的ARMA-GARCH模型预测效果更好,预测相对误差仅为1.3%,可为相关投资者提供参考依据.Data of the Shanghai daily closing index between December 20, 2000 and June 20, 2016 is used for modeling and short-term prediction analysis. Firstly, ARIMA model is established. Then, due to the presence of conditional heteroskedasticity and non-normality of the model residuals, ARMA-GARCH model is established. Distribution types of the residuals norm, GED and t are taken into account. By comparison of prediction accuracy, it is found that the t-ARMA-GARCH model has better prediction effect, the relative error of prediction is only 1. 3%, which can provide helpful suggestions for relevant investors.

关 键 词：上证综合指数 ARIMA模型 t-ARMA-GARCH模型 预测 

分 类 号：F224.7[经济管理—国民经济]