基于持续期模型对高频金融数据分析  

High Frequency Financial Data Analysis Based on Duration Model

作　　者：钱有程[1] 王暘[2] 

机构地区：[1]吉林化工学院理学院,吉林吉林132022 [2]商务部国际贸易经济合作研究院,北京100710

出　　处：《吉林化工学院学报》2014年第9期81-83,共3页Journal of Jilin Institute of Chemical Technology

摘　　要：本文意在研究高频金融数据具有的性质和特点以及时间序列持续期模型的适用性.利用中国石油2014年6月9日至6月18日8个交易日1分钟高频交易数据,用时间序列持续期模型进行分析,得到交易的相互依赖现象,说明股票交易期间的具有聚集效应.这说明短时间间隔伴随着短交易时间,长时间间隔伴随长交易时间.同时也说明股票交易具有间歇性频繁、平淡,也验证了持续期模型对研究高频数据的特性的合理性.This paper intended to study the nature and characteristics of high frequency financial data and application of time series duration model. Based on duration model of time series analysis analyses the data which gathered the one minute high frequency financial data of 8 trading day of Petro China from June 9th,2014 to June 18th. It get mutual dependence transaction duration,i, e. stock transactions continued accumulation period. It shows that the short transaction time interval behind often with short transaction time interval, and long transaction time interval behind often with long transaction time interval. At the same time also shows that in our country ＇s stock market,often trading over a period of time is very frequent, and trading in a period of time is very dull. It is also available to study the on high frequency data with duration model.

关 键 词：持续期模型 Ljung-Box统计量 高频金融数据 

分 类 号：O212.1[理学—概率论与数理统计]