Aerospace Instrument-Making Annotation << Back Bayesian estimation by using neural network O.A. Stepanov, O.S. Amosov The relation between the traditional (minimum variance) algorithms for the random vectors estimation and the algorithms based on the use of neural networks is investigated. It is shown, in particular, that linear neural networks at the appropriate choice of the criterion used for its preliminary training, provide estimates with similar properties as at the Bayesian and neural network algorithms. The derived results are discussed. The examples are considered.