The implementation of a generalized cross validation algorithm using deflation techniques for linear systems (Q1917400)

The main problem is the application of computational deflation techniques for solving linear systems. Adaptation of general deflation techniques is considered with respect to applications in generalized cross validation function minimization. The deflation techniques described in the paper are parallelized and some timing runs are performed with different size of data sets on a Cray Y-MP 2D computer. A comparison with a standard routine based on LU factorization and reverse Gauss-Seidel deflation techniques with an underlying Jacobi iteration is performed. The results of this work are presented and the efficiency of the proposed techniques is illustrated.