Petrov K. Comparative structurally-parametric identification of the models of scalar multifactor estimation

Dissertation is devoted to research and development general methodology of decision a problem of formal structurally-parametric identification the multifactor estimation models within the framework of a decision making theory, which is based on the use of ideas of comparative identification theory. The model of multifactor estimation invariant to the dimension of factors, characterizing alternatives in the class of Kolmogorov-Gabor polynomials is synthesized. The methods of comparative parametric and structurally-parametric identification of a estimation models are developed on the basis of qualitative expert information about the relative preference of alternatives. The methods of determination the generalized multifactor estimations of alternatives are developed in the conditions of interval uncertainty parameters of models of estimation and factors, characterizing alternatives; methods of ranging and extreme alternative determination. The number of the applied problems is solving, particularly, identification of consumer marketing preferences, ordinal classification, social group behavioral management, quality estimation.