A Method of Optimizing Network Topology Structure Combining Viterbi Algorithm and Bayesian Algorithm

Algorithm 1

BPR designing process.

  Input Training samples, number of iterations n, regularization parameters

  Output model parameters

      Initialization parameters

      While i<n do

        according to Maximum a posteriori P(W,H|>u)P(W,H|>u) to solve the model parameters W, H, use the sigmoid function to set the threshold

      End

      Return Recommended list