bei Google ScholarTitelangaben
Larch, Mario ; Tappeiner, Gottfried ; Walde, Janette:
Performance Contest Between MLE and GMM for Huge Spatial Autoregressive Models.
In: Journal of Statistical Computation and Simulation.
Bd. 78
(2008)
Heft 2
.
- S. 151-166.
ISSN 1563-5163
DOI: https://doi.org/10.1080/10629360600954109
Abstract
When using maximum likelihood estimation for spatial models, a well known problem is the computation of the logarithm of the determinant of the Jacobian, especially for problems with a huge number of observation units. In the recent literature there are various promising approaches to account for these numerical difficulties, relying on alternative decompositions or approximations. Recently, a general method of moments approach for estimating these models was developed. We compare all these different approaches with respect to their root mean-squared errors of the estimates and investigate the size and power of hypotheses tests with respect to the spatial correlation and the regression parameters.