Processing approach of non-linear adjustment models in the space of non-linear models

This paper investigates the mathematic features of non-linear models and discusses the processing way of non-linear factors which contributes to the non-linearity of a non-linear model. On the basis of the error definition, this paper puts forward a new adjustment criterion, SGPE. Last, this paper investigates the solution of a non-linear regression model in the non-linear model space and makes the comparison between the estimated values in non-linear model space and those in linear model space.