One-Predictor Regression

Polynomial terms are not the only way to fit a nonlinear shape. In many cases it makes more sense to transform the variable itself directly. Generally this means running a variable through a function before you use it.