A new scheme for calculating weights and describing correlations in nonlinear least-squares fits

摘要: The equations for nonlinear least‐squares analysis are reformulated in terms of dimensionless vectors and matrices. diagonal elements a curvature matrix give the relative weights fit variables. Eigenvectors eigenvalues this used to describe correlations between all parameters, bivariant correlation coefficients may be calculated directly from its elements. With formulation it is easy compare confidence limits, correlations, predictions based on with results Monte Carlo simulations. This provides direct test parabolic approximation. Examples linear biexponential model presented demonstrate these ideas. © 1996 American Institute Physics.