A comparative study for the estimation of parameters in nonlinear models (Q1918308)

The exponential regression is a relatively new noniterative method for estimating the parameters of nonlinear models, proposed by the first author [Appl. Math. Comput. 50, No. 1, 51-58 (1992; Zbl 0746.62061)]. The method of internal regression is another interesting method for estimating the parameters of nonlinear models which does not require the iterations. Based on the simple model \[ y= \theta_1(1- \theta_2 \exp(\theta_3 x)) \] and real data by \textit{H. O. Hartley} [Biometrika 35, 32-45 (1948; MR 10.50)], the authors compare their methods with Gauss-Newton, Newton-Raphson, Marquart and steepest descent methods and offer some general conclusions.