Arkhipova S. Identification of approximate models by a method of a variation of data

The technique of a solution of a task of identification of approximate models optimizing procedures of selection of a structure of a model and choice of the best method of parametrical identification in conditions of a minimum information about input data is offered free from an aspect of distribution of an error. The realization of a technique is based on the algorithms, developed and investigated by the author, of a variation of sample. The index of quality identification exterior in relation to used methods of an estimation is entered. The mode of shaping of a group evaluation of a parameter for an improvement of estimations obtained by different methods is developed.