Estimation of Parameters of Two Non-linear Regression Models Using Assumed Values: Reciprocal Power Regression Models

Abstract

Fitting nonlinear models is not a single-step procedure but it involved a process that requires careful examination of each individual step. Depending on the objective and the application domain, different priorities are set when fitting nonlinear models; these include obtaining acceptable parameter estimates and a good model fit while meeting standard assumptions of statistical models. We propose steps in fitting nonlinear models in this research work. Two reciprocal power regression models were considered with a non-linear data set. Then, the following steps are considered (i) fit the models to the data collected using iterative steps, (ii) to develop a linear model to estimate the parameter β1 and β2 when the initial value (or growth rate β3) lies between -1.0 ≤ β3 ≤1.0 ); using the transform models of the reciprocal power regression model (iii) to find the “best” model between the two models using R2, AIC and BIC. The results show Model B is better than Model A, using the model selection criteria.