ISSN:
1432-1394
Keywords:
Key words. Multivariate Gauss
;
Markoff model
;
Eigenvalue decomposition
;
Principal-component analysis
;
Data reduction
Source:
Springer Online Journal Archives 1860-2000
Topics:
Architecture, Civil Engineering, Surveying
Notes:
Abstract. Geodetic adjustment problems frequently require the solution of large systems of linear equations. An approximation method is presented based on the decomposition of the estimated covariance matrix of the observation matrix, calculated in a pre-processing step, into a system of eigenvalues and eigenvectors. Neglecting the non-dominant eigenvalues and the assigned eigenvectors, the matrix of the residuals is approximated applying the synthesis formula of principal-component analysis. Although the number of observation vectors in the multivariate Gauss–Markoff model is drastically reduced, all unknown parameters are estimated approximately. The described method is tested using a numerical example of satellite altimetry.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/s001900050225
Permalink
