Description: Until now, there is no general methodology for synthesizing non-parametric decisive rules, which requires the search for new approaches to solving this problem and determines the relevance of such research. In real terms, it's necessary to take into account that the signals received in the background of interference, as a rule, have unknown parameters, some of which may be of interest to the consumer. In addition, noise is not necessarily centered noise with an unknown mean value. Therefore, the purpose of the article is to determine the possibilities and ways of applying the informational approach to the problems of synthesizing the decisive rules for detecting and evaluating signal parameters against the background of disturbances with an unknown distribution law. The article generalizes and deploys the ideas of the informational approach to the synthesis of decisive rules for detecting and evaluating signal parameters against the background of additive interferences with an unknown distribution law. New basic theoretical concepts are introduced and on the basis of them formed approaches to the synthesis of nonparametric decisive rules for evaluating discrete and continuous observation parameters. It is proved that the proposed algorithms for detecting and evaluating signal parameters for a Gaussian class and a wide class of non-Gaussian disturbances, whose distribution density can be represented by polygonal models, are completely equivalent to the optimal criterion for maximum probability for parametric algorithms for detecting and estimating quasitermined signal parameters at an apriori unknown power of interference.
Keywords: statistics of relation of informing of observations, principle of a maximum of informing of observations, non-parametric ambiguity, informative measures, informative statistics, asymptotic optimum algorithm of detection of signal, estimation of Fisher’s information