A modelling procedure can be seen as an algorithm that, analyzing a set of data, provides a mathematical description of a phenomenon, which is called a model. In some cases the most suitable description of the data is a dynamical one, and consequently the appropriate class of models to be considered is that of dynamical systems. In this paper the particular case of noise-free data with two-dimensional support is considered. The model class is constituted by systems of two-dimensional difference equations, and corresponds to the family of linear systems as defined in the behavioral approach. In this setup two indices describing the identifiability of the model in the model class are introduced. They provide information on the amount of data that is needed to identify a model in two directions: how many trajectories are needed and how large their support should be.
A behavioral approach to identifiability of 2D scalar systems
ZAMPIERI, SANDRO
1997
Abstract
A modelling procedure can be seen as an algorithm that, analyzing a set of data, provides a mathematical description of a phenomenon, which is called a model. In some cases the most suitable description of the data is a dynamical one, and consequently the appropriate class of models to be considered is that of dynamical systems. In this paper the particular case of noise-free data with two-dimensional support is considered. The model class is constituted by systems of two-dimensional difference equations, and corresponds to the family of linear systems as defined in the behavioral approach. In this setup two indices describing the identifiability of the model in the model class are introduced. They provide information on the amount of data that is needed to identify a model in two directions: how many trajectories are needed and how large their support should be.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.