The art of model fitting to experimental results

Abstract

Model fitting to experimental results is presented within the context of graduate physics laboratories and generalized to other graduate and post-graduate levels, and diverse research fields. In most cases the analysis of experimental results, in terms of mathematical models available to describe the obtained results, extends beyond the numerical minimization of statistical estimators, like the chi-square, in the model's parameter space. Dedicated fitting procedures, not easily or directly available in common data analysis software's packages, are required to obtain the best fitting set of parameters that present a consistent physical meaning. A simple but powerful web-based solution is presented, and its relative advantage in comparison with known commercial and open source solutions is discussed.