Marginal probability distribution determined by the maximum entropy method

We derive a formula for the probability distribution of a random variable if it is statistically dependent on other random variables of a stochastic system supposing that it is in its maximum entropy state. We assume that the conditional probability distribution is given and we use the maximum entropy principle for getting the marginal probability distribution. Such stochastic system may represent, e.g. homogeneous polymer, probabilistic cellular automaton, one-dimensional Ising system, or a string of symbols in communication.