ESTIMATION OF PARAMETERS IN NONLINEAR MODELS: ALGORITHMS AND APPLICATIONS

Oscar Cornejo Zúñiga, Rodrigo Rebolledo Vega

Abstract


This article shows different algorithms for estimating parameters in nonlinear models. They are applied primarily to a database of problems classified as difficult. Later, the article shows the behavior of the algorithms for the study of growth in the anchovy and sardine, and males and females of the common hake by adjusting a Von Bertalanffy model. The Cerrato test is applied for growth comparisons between sexes for the common hake. The algorithms are implemented in a MATLAB environment, showing good behavior regarding CPU time, number of iterations and exactitude in the solution found with respect to certified values of the problems in the database.

 


Keywords


Nonlinear regression; Quasi-Newton methods; Confidence region methods; Growth models

Full Text:

PDF

Refbacks

  • There are currently no refbacks.


UNIVERSIDAD EIA

Sede de Las Palmas: Km 2 + 200 Vía al Aeropuerto José María Córdova Envigado, Colombia. Código Postal: 055428
Tel: (574) 354 90 90. Fax: (574) 386 11 60

Sede de Zúñiga: Calle 25 Sur 42-73 Envigado, Colombia. Código Postal: 055420
Tel: (574) 354 90 90. Fax: (574) 331 34 78
NIT: 890.983.722-6

Sistema OJS - Metabiblioteca |