基于ARCH(q)模型的高频股票交易数据分析  

Empirical analysis of high frequency stock trading data based on ARCH(q) model

作　　者：杨凯 马育欣 张欣然 陈雪妮 田源 王晓红[2] YANG Kai;MA Yu-xin;ZHANG Xin-ran;CHEN Xue-ni;TIAN Yuan;WANG Xiao-hong(School of Mathematics and Statistics,Changchun University of Technology,Changchun 130012,China;College of Mathematics,Jilin Normal University,Siping 136000,China)

机构地区：[1]长春工业大学数学与统计学院,吉林长春130012 [2]吉林师范大学数学学院,吉林四平136000

出　　处：《吉林师范大学学报（自然科学版）》2019年第2期68-72,共5页Journal of Jilin Normal University:Natural Science Edition

基　　金：国家自然科学基金项目(11571051);吉林省教育厅"十三五"科学技术研究规划项目(2016316;JJKH20180767KJ);吉林省大学生创新创业训练计划项目(201810190050)

摘　　要：在大数据背景下,数据的采集和存储技术迅猛发展,人们也越来越关注股市的日内高频交易规律.本文基于ARCH(q)模型对中国国贸(600007)股票的高频日内交易数据进行了实证分析,检验了数据的平稳性和异方差性,研究了模型的定阶问题和预测问题,并与传统预测方法进行了比较.拟合结果表明,本文所采用的方法较传统的方差齐性假设下所算得的置信区间更加精确.Under the background of big data,the data collection and storage technology is developing rapidly.People pay more and more attentions to the law of high frequency trading in the stock market.Based on ARCH ( q ) model,this paper made an empirical analysis of the high-frequency intraday trading data of China International Trade (600007) stocks,test the stationarity and heteroscedasticity of the data,and studied the problems of model ordering and prediction,and compared with traditional forecasting methods.The fitting results show that the method used in this paper is more accurate than the traditional method under the assumption of homogeneity of variance in terms of the confidence intervals.

关 键 词：ARCH 模型 高频数据 预测 

分 类 号：O212[理学—概率论与数理统计]