ABSTRACT

A fact that stands out distinctly in Chapters 1 and 2 is that the model is both the end-product and the centerpiece of every identification exercise. A good model necessarily requires “good” quality data and the choice of a good starting model. While generating quality data is achieved by a proper input design, the choice of the initial model structure depends on the a priori knowledge of the suitable model type and structure. Usually the situation is that we begin with a good guess and improve upon the initial choice if necessary (as in the case study of Chapter 2). The final decision is governed by how well the model suits the purpose, cost (of estimation) considerations, the merits and limitations of a given model and other constraints.