ISBN:  Unavailable

Category: Unavailable

Page count: 19

Abstract: "A version of parallel computation in dynamic programming for optimal string-to-network alignments is discussed. The specific application is pattern analysis in which Markov networks are used as models of classes of strings. The dynamic programming computes maximum probability string-to-network alignments with a cost function based on relative frequencies. The concurrent-processor computer is a Sequent Balance, a machine with processors, memory modules, and I/O controllers connected to a single, high-speed bus. The implementation and experimental results are described."