Advances and Applications in Statistics
Volume 55, Issue 2, Pages 253 - 268
(April 2019) http://dx.doi.org/10.17654/AS055020253 |
|
OUTLIER DETECTION IN LINEAR TIME SERIES REGRESSION MODELS
Arfa Maqsood, S. M. Aqil Burney, Suboohi Safdar and Tahseen Ahmed Jilani
|
Abstract: In this article, the problem of detecting and identifying outliers in a univariate time series model is addressed. We consider the univariate autoregressive (AR) process when the observations are perturbed by different kinds of outliers. We employ the sequential test and modified sequential test for detecting the positions and types of outliers. These tests are applied iteratively to the new set of residuals until no outlier is obtained. The methods are simply illustrated with simulated examples and also demonstrated by analyzing a real time series data. |
Keywords and phrases: outliers, univariate time series, autoregressive process, sequential test, modified sequential test.
|
|
Number of Downloads: 305 | Number of Views: 944 |
|