Method of full sorting out in the problem of regressive polifactor-analisis

The paper is considers the problem of the seguential complication of the regressive polyfactor-model with full sorting out of all the possible factors combinations at their given number. There is offered the program-realized algorithm of full sorting out for the construction of the best model with the fixed number of factors, that permits to select an optimal model from the set of the best models for the different number of factors according to the well-known criteria of significans.