Abstract:
In this paper a new characterization for the quality of stochastic, linear-in-parameters, difference equations models is given. This requires a good model to have accurat...Show MoreMetadata
Abstract:
In this paper a new characterization for the quality of stochastic, linear-in-parameters, difference equations models is given. This requires a good model to have accurate parameter estimates and uncorrelated residues. A criterion based on this characterization is derived and compared with other criteria which depend on minimizing the mean square of model residues.
Published in: IEEE Transactions on Systems, Man, and Cybernetics ( Volume: 10, Issue: 2, February 1980)