Nonlinear Regression Modeling of Data Employing the Gauss-Newton Method

Date of Graduation

Fall 2006

Degree

Master of Science in Mathematics

Department

Mathematics

Committee Chair

George Mathew

Abstract

Linear regression models are useful for estimating statistical relationship between related variables of a data set. In many practical situations a linear regression model is not an appropriate fit for the data. In this thesis, we analyze data by nonlinear regression methods to determine an appropriate relation between the variables involved. The estimation of the parameters involved in the nonlinear regression model is nearly impossible by any closed-form solution approached. The modeling is done by iteration procedures involving the Gauss-Newton method. A comparison of the results is also provided.

Keywords

error sum of squares, Gauss-Newton method, linear regression, nonlinear regression, residual plot analysis

Subject Categories

Mathematics

Copyright

© Michael William Edwards

Citation-only

Dissertation/Thesis

Share

COinS