Journal Volumes


Visitors
ALL : 900,872
TODAY : 1,088
ONLINE : 67



















  JOURNAL DETAIL



Effect of Preliminary Unit Root Tests on Prediction Intervals for Gaussian Autoregressive Process with Additive Outliers


Paper Type 
Contributed Paper
Title 
Effect of Preliminary Unit Root Tests on Prediction Intervals for Gaussian Autoregressive Process with Additive Outliers
Author 
Wararit Panichkitkosolkul and Sa-aat Niwitpong
Email 
snw@kmutnb.ac.th
Abstract:

The  preliminary  unit  root  test  has  been found  to be  a  useful tool  for  improving  the  accuracy  of a  one-stop-ahead  predictor  and  prediction  interval  for  the first-order  autoregressive  process  when  an  autoregressive  coefficient  is  close  to  one. This paper  applies  the  aforementioned  concepts  of the  preliminary  unit root  test  in  order  to  improve the efficiency  of  prediction  intervals  for  the  Gaussian  autoregressive  process  with additive  outliers. The  preliminary unit  root  tests  considered  are  the  augmented  Dickey-Fuller (ADF) test  and  the  Shin  et  al.’s (SSL)  test. In  addition, the  analytic  expressions  of  the  coverage  probability  of  prediction  intervals  are  derived, and the  structure  of  the coverage  probability  was  proved  to  be  independent  from  the mean of  the  process  and the  parameter  of  the  innovation, it  is  a function  of  the  autoregressive  coefficients  only. For  the  parameter  estimation  of processes  we  use  the generalized  M-estimates. The  coverage  probabilities  and the widths  of  the  standard  prediction  interval, the prediction  interval  following  the ADF  test, and  the prediction  interval  following  the  SSL  test  are  also  compared  via  simulation  studies. Simulation  results  have shown  that the SSL  test  can help  to  improve  the accuracy  of  the  prediction  intervals  with  additive  outliers, especially  when  the  sample  size is  large. The  performance  of the  proposed  prediction  intervals  is  illustrated  with  an  empirical  application.

 

Start & End Page 
8 - 29
Received Date 
2011-05-25
Revised Date 
Accepted Date 
2011-09-30
Full Text 
  Download
Keyword 
additive outlier, autoregressive process, prediction interval, preliminary unit root test
Volume 
Vol.39 No.1 (JANUARY 2012)
DOI 
SDGs
View:488 Download:131

Search in this journal


Document Search


Author Search

A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

Popular Search






Chiang Mai Journal of Science

Faculty of Science, Chiang Mai University
239 Huaykaew Road, Tumbol Suthep, Amphur Muang, Chiang Mai 50200 THAILAND
Tel: +6653-943-467




Faculty of Science,
Chiang Mai University




EMAIL
cmjs@cmu.ac.th




Copyrights © Since 2021 All Rights Reserved by Chiang Mai Journal of Science