ORDER SELECTION OF AUTOREGRESSIVE MODELS

被引:29
作者
DJURIC, PM [1 ]
KAY, SM [1 ]
机构
[1] UNIV RHODE ISL,DEPT ELECT ENGN,KINGSTON,RI 02881
关键词
D O I
10.1109/78.165674
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
This correspondence addresses the problem of order determination of autoregressive models by Bayesian predictive densities. A criterion is derived employing noninformative prior densities of the model parameters. The form of the obtained criterion coincides with that of Rissanen in [16]. Simulation results are presented which demonstrate the good performance of the criterion, and comparisons with four other popular approaches verify its superiority in many cases.
引用
收藏
页码:2829 / 2833
页数:5
相关论文
共 19 条
[1]   GOODNESS OF PREDICTION FIT [J].
AITCHISON, J .
BIOMETRIKA, 1975, 62 (03) :547-554
[2]  
AITCHISON J, 1975, STATISTICAL PREDICTI
[3]   NEW LOOK AT STATISTICAL-MODEL IDENTIFICATION [J].
AKAIKE, H .
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1974, AC19 (06) :716-723
[4]  
[Anonymous], 1983, LEAST ABSOLUTE DEVIA
[5]  
[Anonymous], 1988, MODERN SPECTRAL ESTI
[6]  
Box G.E.P., 1992, BAYESIAN INFERENCE S
[7]  
BOX GEP, 1970, TIME SERIES ANAL FOR
[8]   STATISTICAL-THEORY - THE PREQUENTIAL APPROACH [J].
DAWID, AP .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES A-STATISTICS IN SOCIETY, 1984, 147 :278-292
[9]  
DJURIC PM, 1990, P IEEE ICASSP, P2415
[10]   LATTICE FILTERS FOR ADAPTIVE PROCESSING [J].
FRIEDLANDER, B .
PROCEEDINGS OF THE IEEE, 1982, 70 (08) :829-867