A Comparison of Methods Fitting von Bertalanffy Growth Curve
摘要
To compare the methods of fitting von Bertalanffy growth curve, three methods, namely, Ford-Walford-Chapman, Beverton, and Gauss-Newton method were first reviewed and three data sets: generated data, abbreviated generated data, and observed data were tested to compare the goodness of fit of these method. The criteria for the goodness of fit were Index of Fit, Sum of Square for Error, Average Sum of Square for Error, and Coefficient of Determination. The results showed that the Gauss-Newton method had the best fit among methods tested. The parameters estimated by Ford-Walford-Chapman method were acceptable if the coverage of samples was large and intensive. Beverton method was actually a non-scientific 'trial and error' method with no promise of obtaining the best fit. Discussion was also made along the statistical aspects.