ITEMS

modern scholarly publishers in the finest tradition

Journals

	Search

Journal of Automation and Information Sciences

ISSN for PRINT: 1064-2315

Institutional price:	$2573.00
Issues per year:	12

Best Paper Award Selection - Editorial Board Site

2003, Volume35

63 pages

DOI: 10.1615/JAutomatInfScien.v35.i1

Issue price - $221.00

The Use of Fuzzy A Priori Information for Estimation of Regression Parameters

Arnold S. Korkhin
Research Institute of Ferrous Metallurgy Automation, Dnepropetrovsk, Ukraine

Victor N. Mizernyi
Vinnitsa National Technical University, Ukraine

ABSTRACT

A method of a priori information consideration, based on description of definition domain of regression parameters by equality constraints and by inequality constraints, where either right-hand or left-hand parts of constraints are prescribed fuzzy, is suggested. Statistical properties of estimates are studied. The described technique is applied for derivation of models of non-stationary stochastic systems, which can be reduced to a model of regression with switchings.