Adaptive Systems

Identification algorithms in regression type problems with zero median noise

P. I. Balk

Irkutsk

Abstract: The paper is concerned with parametric identification of linear plants represented as a regression model. Algorithms are obtained which should not satisfy the assumption of zero mean noise, a typical requirement for a priori specification of expectation. The underlying conception of the algorithms is common in theory of statistical solutions, viz. that optimization problems are regarded with the minimal empirical estimate of the mean risk as the criterion.

UDC: 62-50:519.2


Received: 12.01.1978

 English version: Automation and Remote Control, 1979, 39:12, 1793–1803

Bibliographic databases:

•