Computationally efficient methods for estimating the updated-observations SUR models

Computational strategies for estimating the seemingly unrelated regressions model after been updated with new observations are proposed. A sequential block algorithm based on orthogonal transformations and rich in BLAS-3 operations is proposed. It exploits efficiently the sparse structure of the data matrix and the Cholesky factor of the variance–covariance matrix. A parallel version of the new estimation algorithms for two important classes of models is considered. The parallel algorithm utilizes an efficient distribution of the matrices over the processors and has low inter-processor communication. Theoretical and experimental results are presented and analyzed. The parallel algorithm is found for these classes of models to be scalable and efficient.