15. Metrics and evaluation of machine learning models

Now that we can fit some basic models, how do we know which is best? How do we know how good of a model it is? What metrics are available to use? R^2, MAE, MSE, RMSE and more are available to quantify error in models.

0:00 how good is your model?
0:56 R^2 correlation coefficient for linear models
3:30 why R^2 doesn't work for non-linear models
7:34 alternative metrics MAE, MSE

[richardaversa]

Great caution about using R² for nonlinear models

[evansdiegbe]

Can I get your mail to send you a question for evaluation of ml model