Abstract:
The paper is concerned with identification in terms of criteria other than minimal r.m.s. error. Vector identification criteria are studied. Use of dispersional functions permits extending the approach to nonlinear plants and leads to a plant model even when correlation methods are inapplicable.