A two-stage fuzzy least squares procedure for fitting
240 / 125
Abstract
In this paper, fitting of linearized von Bertalanffy growth model is revisited when response variabe is reported in intervals
corresponding to various values of explanatory variable. Earlier, Tanka’s linear programming methodology was used to
handle this problem. However, this approach has several limitations, which are highlighted. Accordingly, a more efficient
two stage procedure based on fuzzy least squares, is employed. The methodology is thoroughly discussed and, for its application,
relevant computer programs are developed in “Nonlinear programming solver LINGO, Version 8†software package. Finally,
an illustration to pearl oyster age-length data is discussed.