On the identification of large multilinear systems (Q2463652)

The authors deal with bilinear or high-order multilinear and multivariate time series which are controlled by a series of also multivariate input variables. Such stochastic processes can be represented as the output of a noisy bilinear system. An algorithm is described which allows the identification of the system structure. The performance of the algorithm is illustrated by a Monte Carlo study. It is also discussed how the considered models are related to bilinear systems in state space form and their input-output realizations.