Question

i have about 5 different regression models including non linear. do I choose the best one by looking at the plot or only by comparing R^2?

Answer 1

It is given that there five different regression models, including nonlinear.

Plots are useful to describe whether the fitted model is linear or nonlinear.

R-square (R2) is useful to explain the preciseness of the model. If the model (linear or nonlinear) has an R-square value closest to 1, then it can be said that the fitted regression model is best.

Therefore, the appropriate method to choose the best regression model from the five different regression models, including nonlinear is R-square.