
Accession Number : AD0739966
Title : Least Squares Estimation for a Class of NonLinear Models.
Descriptive Note : Technical rept.,
Corporate Author : WISCONSIN UNIV MADISON DEPT OF STATISTICS
Personal Author(s) : Guttman,Irwin ; Pereyra,Victor ; Scolnik,Hugo D.
Report Date : SEP 1971
Pagination or Media Count : 23
Abstract : A new method for determining least squares estimators for certain classes of nonlinear models is discussed. The method is an extension of a variable projection method of Scolnik (1970), and involves the minimization of a modified functional. The feature of minimizing this modified functional is that for a certain class of nonlinear models, called the constant coefficients case, only one half the parameters are involved initially. To find the estimators of the remaining parameters is straight forward and relatively easy. This new twostep procedure is shown to be equivalent to the overall least squares procedure. The authors also discuss the case of a class of models called the variable coefficients class. For this case, the authors formulate a new algorithm for determining the estimators which make use of approximate confidence regions for the parameters. (Author)
Descriptors : (*STATISTICAL ANALYSIS, MATHEMATICAL PREDICTION), LEAST SQUARES METHOD, CONFIDENCE LIMITS, PROBABILITY, THEOREMS
Subject Categories : Statistics and Probability
Distribution Statement : APPROVED FOR PUBLIC RELEASE