摘要
应用分数差分自回归滑动平均模型即ARFIMA模型,研究金融时间序列.通过对外汇市场日元/人民币汇率数据的实证分析,得出结论:汇率日收益率序列不存在长期相关性,而汇率日绝对收益率序列存在显著的长记忆性.
In this paper, ARFIMA（p,d,q） model was used to study financial time series. The study of the exchage rate data of JPY/RMB in Chinese foreign exchange market showed that data of daily exchange rate yield don＇t exsist long correlation, while the data of daily absolute exchange rate yield exsist significant long-term dependence.
出处
《海南师范大学学报（自然科学版）》
CAS
2008年第4期448-450,共3页
Journal of Hainan Normal University（Natural Science）
关键词
汇率
长期相关
长记忆模型
exchange rate
long-term dependence
long-term model
