In the paper general approach is considered to the synthesis of the adapted moment criterion of quality of testing statistical hypotheses, based on the use of stochastic polynomials and moment-kumulant description of random values. It is shown that the use of joint moments of different orders is given by possibility to take into account cross-correlation properties of random values and their non-Gaussian distribution. The mathematical results allows to synthesize the new algorithms of signal detection and recognition on a background the correlated non-Gaussian noise with the best high-quality propertie .
testing of statistical hypotheses, moment-kumulant description, non-Gaussian random value, correlation
"Momentnыi kryteryi kachestva proverky statystycheskykh hypotez dlia obrabotky syhnalov na fone korrelyrovannыkh nehaussovskykh pomekh" ,
Information Processing Systems,