Representation and identification of non-parametric nonlinear systems of short term memory and low degree of interaction

In this paper, we propose a new representation which is particularly useful for a class of non-parametric nonlinear systems that have short term memory and low degree of interaction. Advantages and disadvantages of this representation are discussed and compared to existing methods both theoretically and numerically. Furthermore, results regarding structural estimation based on the analysis of variance and on full scale identification are also provided.