Multiple Choice Flashcards

They replace data validation, residual diagnostics, and out-of-sample model checking They directly report classification accuracy They are most useful for comparing candidate models fit on a common basis They prove the fitted model is causal and free of confounding