Bayesian analysis of non-negative ar(2) processes |
| |
| Authors: | Jiri Andel Manuel Garrido |
| |
| Institution: | 1. Department of Statistics , Cjarles University , Sokolovska 83, Prague 8, Czechoslovakia, 18600;2. Campus de Montegancedo , Universidad Politecnica , Boadilla del Monte, Madrid, Espana, 28660 |
| |
| Abstract: | A non-negative AR(2) process with exponentially distributed white noise is investigated in the paper. It is assumed that the autoregressive parameters are random variables with a vague prior density. They can be esto,ated by their posterior expectations. Explicit formulas for these estimators are derived and their strong consistency is proved. An approximation to the estimators is proposed which is easier for calculation. The results are illustrated in a simulation study |
| |
| Keywords: | Autoregressive process BAYESian approach non-negative veriables strongly consistent |
|
|
