Декодирование последовательности состояний бинарной скрытой марковской модели, представленной в виде алгебраической байесовской сети, по последовательности наблюдений

Probabilistic graphical models class including hidden Markov models and Bayesian networks proved to grant effective technique for representation of uncertainty in knowledge with actively developing theoretical and algorithmic apparatus; such models found many applications in the fields of speech recognition, signal processing, bioinformatics, natural language processing, digital forensics etc. The paper suggests a decoding algorithm for hidden states of binary linear hidden Markov models represented in the form of algebraic Bayesian networks; its correctness is proved. The presented algorithm completes the set of methods of such models.