ASSUMPTION EVALUATION IN APPROXIMATE MODELING

System identification can be considered as an approximate modelling problem on the basis of observed data and a priori knowledge. Assumptions are often introduced on the set of the considered models and on the characteristics of the noise affecting the data; the validity of such assumptions must then be evaluated on the basis of the identified model family. This paper refers to the Frisch scheme applied to dynamic system identification and introduces a procedure to evaluate the violation of the assumptions underlying the considered scheme.