Data-Fitted Models

In this chapter we consider the most common setting for model discrimination, where the proposed model functions, sometimes referred to as the deterministic part of the modeling scheme, contain unknown parameters. These models, since they are estimated through parameter estimates based on the responses, are called data-fitted models. Unlike the completely determined models, which occur mainly in science or engineering contexts, these models are used in both the hard and soft sciences. Based on empirical data, the model's unknown parameters are estimated using the method of least squares. Substitution of the estimates for the models in S leads to k estimated or fitted models. In this setting we 56search for efficient DMs and DFs to apply to the set of competing models.