In order to estimate autoregressive moving average models with a general polynomial time trend, the restricted or residual likelihood is considered. The autoregressive moving average parameter estimator maximizing the restricted likelihood has shown to have the second-order bias equivalent to the maximum likelihood estimator computed under the assumption that the polynomial time trend parameters are known.
Bibliographical noteFunding Information:
This research is supported by a grant from the SRCCS of the Korean Science and Engineering Foundation.
- Autoregressive moving average process
- Polynomial time trend
- Restricted likelihood
- Restricted maximum likelihood estimator