| Article ID: | iaor19941034 |
| Country: | Italy |
| Volume: | 22 |
| Issue: | 64 |
| Start Page Number: | 43 |
| End Page Number: | 55 |
| Publication Date: | Dec 1992 |
| Journal: | Ricerca Operativa |
| Authors: | Spedicato Emilio, Bodon E. |
| Keywords: | statistics: regression |
Three different versions of the implicit QR algorithm in the ABS class of methods applied to least squares problems are experimented. The results show the accuracy of the algorithms in comparison with the standard QR orthogonalization algorithm and the standard SVD decomposition algorithm based on the Householder transformations.