B. F. Grin'ko

The article presents data, revealing the main points of methodology of estimation of parameters of false information at output of complex high-responsibility systems from results of observation at a confined time interval. Presented methodology is based on use of methods of short samples processing, when amount of observation data is not sufficient for usual statistical methods. Analytical relationships for calculation frequency of forming false information are presented for the case when the process of forming can be described by piecewise linear Markov processes. Possible schemes of development of used mathematical models and obtained estimates of parameters of false alarms at the output of the considered system are proposed. Requirements to the used mathematical models, providing "the best" model for chosen class of distributions are determined. Concepts of guaranteed approach and guaranteed estimate according to the considered case are introduced. Proposed approach in general is applicable to a wide class of explored information systems, processes, and phenomena