Abstract:
The paper presents an algorithm for evidence propagation in algebraic Bayesian networks in order to calculate a posteriori probabilistic estimates of network elements. The implicit assumptions that guarantee the correctness of the a posteriori inference algorithm results are discovered and considered.