ABSTRACT
One important assumption in a linear regression model is that the relationship between the response variable and predictor variables is linear. When the relationship is not linear, we often must transform either the response variable or the predictor variable or both. Transformation for linearity is often a necessary step of a modeling project. In many situations we have subject matter knowledge that provides the basis for selecting a specific model form, which may be impossible to transform to linear. Nonlinear regression using the least squares concept is often used for estimating model coefficients.
