Standardized Comparison of Nonlinear Model Fitting Algorithms
S. LeRoux, Antoine Messéan, Jean-Pierre Vila · 1982
An orthogonality true convergence criterion has been implemented in several well-known nonlinear least squares algorithms. This authorized a normalized comparison of these algorithms on several typical nonlinear models. A strategy of use of this criterion is proposed to improve convergence process, classically based on relative change termination criteria.