A Front Line algorithm for error estimation in data sets with nonuniform sampling distribution

A new method for the heuristic error estimation, based on nonuniform sampling, is presented. The error is defined as a maximal allowable change of the particular parameter implying small (preset) change of the objective function value. The method is suited to the sampling obtained from a genetic algorithm. The study case is shown. Application to Coulomb Excitation experiments (nuclear physics) data analysis in multidimensional space is outlined.