Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back Testing of parametrical hypotheses at different supervision intervals of sampled values S.V. Vlasova The question about parameter assessment when part of the experimental data is beyond under control in the time interval T is considered in the article. The suggested approach is based on the application of the technique of testing parametrical hypotheses. For this method, however, it is necessary to identify the analytical form for distribution law of observed random variable. This task could be solved by using conditional density function for sampled values, the condition for which will be the number of observed sampled values. The fact that the maximum likelihood estimation in such cases can give less proper result than the offered approach is demonstrated in the paper. The boundary of critical region at the process of accepting competing parametrical hypotheses using the Pirson theorem is defined by the constant C, which could be chosen on the basis of variational principals with application of optimum criteria. Thus it allows us to take into account the consequences of accepted decision, which can have different weighs (costs) of errors. Keywords: parametrical hypothesis, parameter assessment, optimum criterion for splitting sampled area, functional, variation of functional. Contacts: hpric@mail.ru Pp. 36-40.