[1]洪兆萍,杜秀丽.ARFIMA模型参数贝叶斯估计的渐近性质(英文)[J].南京师范大学学报(自然科学版),2008,31(02):16-22.
 Hong Zhaoping,Du Xiuli.Asymptotic Properties of Parametric Bayesian Estimation in ARFIMA Models[J].Journal of Nanjing Normal University(Natural Science Edition),2008,31(02):16-22.
点击复制

ARFIMA模型参数贝叶斯估计的渐近性质(英文)()
分享到:

《南京师范大学学报》(自然科学版)[ISSN:1001-4616/CN:32-1239/N]

卷:
第31卷
期数:
2008年02期
页码:
16-22
栏目:
数学
出版日期:
2008-06-30

文章信息/Info

Title:
Asymptotic Properties of Parametric Bayesian Estimation in ARFIMA Models
作者:
洪兆萍;杜秀丽;
南京师范大学数学与计算机科学学院, 江苏南京210046
Author(s):
Hong ZhaopingDu Xiuli
School of Mathematics and Computer Science,Nanjing Normal University,Nanjing 210046,China
关键词:
贝叶斯方法 ARFIMA模型 后验分布 渐近性质
Keywords:
Bayes ian m ethods ARFIMA models poster io r d istr ibution asym ptotic prope rties
分类号:
O212.8
摘要:
首先根据贝叶斯定理得到ARFIMA模型参数的后验边缘分布,并选择后验边缘分布的众数作为参数的估计值.参照季节性ARFIMA模型的极大似然估计的渐近性质的证明思路,证明了模型参数的贝叶斯估计具有相合性、有效性和渐近正态性.最后,对参数的贝叶斯估计方法的大样本性质进行仿真模拟,结果表明当时间序列样本足够大时,参数的估计值越来越接近于真实值.
Abstract:
Them arg inal po ster io r distr ibution of the param eter in theARFIMA m ode ls is presen ted by Bayes theo rem and the mode o f them arg inal posterior d istr ibu tion is choosed as the estim ator. Then, fo llow ed the analysis of the asympoto tic properties of max imum likelihood estim a tion fo r the seasonal ARFIMA m ode ls, the cons istency, e fficiency and asympto-t ic norma lity o f the Bayesian estima tor are prov ed. Fina lly, large sam ple perform ance o f the Bayesian estim ates is ex amined by sim ulations. It is shown that the estim ates behave we ll when the sam ple size is large enough.

参考文献/References:

[ 1] Granger C W J, Joyeux R. An introduction to long-m em ory tim e series m ode ls and fractional d ifferenc ing [ J]. Journa l o f T im e Ser ies Ana lys is, 1980, 1( 1): 15-39.
[ 2] Chan N H. Estim a tion of long-m em ory tim e series: a survey of like lihood-based me thods[ J]. Advances in Econom etr ics,2006, 20( 2) : 89-121.
[ 3] Je ffrey S Pa,i Na lin iR. Bayes ian ana ly sis o f autoregressiv e frac tiona lly in integrated mov ing-averag e processes[ J]. Journal o f T im e Ser ies Ana lys is, 1995, 19( 1) : 99-111.
[ 4] H annan E J. The asym ptotic theo ry o f linear tim e-series m ode ls[ J]. Journal of App lied Probab ility, 1973, 10( 4): 913.
[ 5] Chang Xuejiang, ChenM ing, Wang M ingsheng. T im e Ser ies Ana lysis[M ]. Be ijing: Ch inaH igher Education Press, 1993.
[ 6] Pa lm aW, Chan N H. Effic ient estima tion o f seasona l long-dependent processes[ J]. Journal of T im e Series Ana lysis, 2005,26( 6): 864-891.
[ 7] Dahlhaus R. E fficient param eter estim ation for se l-f sim ilar pro cesses[ J]. The Anna ls of Statistics, 1989, 17( 4) : 1 749-1 766.
[ 8] Tan iguch iM, Kakizaw a Y. Asym pto tic Theory of Statistical In ference fo r T im e Ser ies[M ]. New York: Spr ing er-Ve rlag,2000.
[ 9] G raybiu F A. M atr ices w ith Applications in Sta tistics[M ]. 2nd ed. B elm ont, CA: W adsWo rth, 1983.

备注/Memo

备注/Memo:
Corresponding autho r: H ong Zhaop ing, teach ing ass istan t, m ajored in fin ancia l statistics. E-m ail:zphong2002@ yahoo. com. cn
更新日期/Last Update: 2013-05-05