ABSTRACT
Regression (curve fitting) is a technique for developing a possible mathematical relationship between the dependent variable (Yi) and one or more independent variables. There are several ways to categorize or differentiate models, one being steady-state in contrast to dynamic (or transient). Dynamic (transient) models have a time-dependent response. There are several standard hypotheses concerning the simple linear regression model. The objective of the model is to fit the natural phenomena, and not fit the noise or experimental error on the data, to remove all variation due to the independent variable. The independent variables might truly be separate variables (like temperature, composition, age, size, etc.), or they could be nonlinear transforms of one or a combination of variables. The use of powers or functional transformations could make the magnitudes of terms very different. Linear regression means that the coefficients in the model appear linearly.
