Matrix computations (3rd ed.)
Matrix algorithms
Hi-index | 0.00 |
In this paper we analyze parallel processing in clusters of computers of an improved prediction method based on RBF neural networks and matrix decomposition techniques (SVD and QR-cp). Parallel processing is required because of the extensive computation found in sucn an hybrid prediction technique, the reward being better prediction performance and also less network complexity. We discuss two alternatives of concurrency: parallel implementation of the prediction procedure over the ScaLAPACK suite, and the formulation of another parallel routine customized to a higher degree for better performance in the case of the QR-cp procedure.