when the model is linear in the coefficients, they can be estimated by a procedure called linear regression. If the model is nonlinear in the coefficients, estimating them is referred to as nonlinear regression. In either case, the simplest adequate model (with the fewest number of coefficients) should be used.