您的位置:首页  > 论文页面

人民币汇率的柯西型自相关模型描述

发表时间:2015-09-30  浏览量:1392  下载量:412
全部作者: 李园春,李明
作者单位: 华东师范大学信息科学技术学院
摘 要: 汇率因受货币政策、国际收支等因素的影响而显示出较强的复杂性,自相关函数(autocorrelation function,ACF)会反应出汇率的重要信息。这里以人民币兑美元、港元、日元3种汇率为代表,以1998年至2014年人民币汇率中间价的部分数据为研究对象,计算得到了汇率数据的ACF. 以拟合曲线与原曲线的均方误差(mean square errors,MSE)为衡量标准,用柯西型过程对3种汇率的ACF进行拟合。结果表明,人民币汇率ACF衰减较慢,具有幂律分布特性,可以用柯西型ACF很好地描述。根据拟合结果计算出的3种汇率的Hurst指数均接近于1,反映出汇率数据强长相关性的特点,表明汇率分析和预测的难度。
关 键 词: 信息处理技术;自相关函数;柯西型过程;汇率;长相关
Title: Description of RMB exchange rate using autocorrelation function model of Cauchy type
Author: LI Yuanchun,LI Ming
Organization: School of Information Science Technology, East China Normal University
Abstract: Exchange rate is complex due to many factors such as monetary policies and balance of international payments. Autocorrelation function (ACF) can help to research exchange rate. Middle prices of RMB exchange rate against USD, HKD, and JPY from the year 1998 to 2014 were selected as research objects. The ACF curves of these three exchange rates were obtained and fitted by Cauchy-class process respectively, measured by mean square errors (MSE) between fitted curve and corresponding original curve. The results show that the ACP curves of RMB exchange rate decay slowly with characteristics of power law, and can be described satisfactorily by Cauchy-class process. Based on the results of curve fitting, Hurst parameters were obtained. The three Hurst parameters, all close to 1, which reveal that exchange rate has very strong long-range persistence and is hard to be analyzed and forecasted.
Key words: information processing technology; autocorrelation function; Cauchy-class process; exchange rate; long-range dependence
发表期数: 2015年9月第18期
引用格式: 李园春,李明. 人民币汇率的柯西型自相关模型描述[J]. 中国科技论文在线精品论文,2015,8(18):1910-1916.
 
0 评论数 0
暂无评论
友情链接