Applying Monte Carlo methods to analysis of nonlinear regression models

摘要

This paper presents a criterion, called coefficient stability, for determining errors in the coefficients of nonlinear regression models describing inexact data. A method for the estimation of coefficient stability is also described. The criterion is illustrated by a computational experiment with data obtained by measuring the refractive index as a function of the wavelength in the 400–1000 nm band for a transparent polymer. The convergence of the criterion to a known analytical solution for the case of linear regression is also studied.