Анотація:
Important inequality is shown out in relation to Shennon mutual information on two dependent boole casual sizes. It is used for proof of estimations from below in the theory of complication oftasks of optimization and Bayesian pattern recognition procedures.