Developing Systems

Linear regression with restrictions on the coefficients of correlation

Yu. M. Tsodikov

Moscow

Abstract: The parameters of a linear process model are determined from a regression equation with prior data on the model input and output correlation coefficients represented as the range of possible coefficients. The solution is reducible to quadratic programming.

UDC: 62-505.72, 519.272


Received: 24.12.1984

 English version: Automation and Remote Control, 1986, 47, 548–552

Bibliographic databases:

• 
•