A linear baseline underfits both the training and validation sets on a complex nonlinear problem. Which change is most likely to reduce bias and improve fit, assuming enough data is available?
Answer & explanation
Correct answer: C
WHY C: When both training and validation performance are poor, the model likely has high bias, and a more flexible model can reduce that bias. WHY NOT A: Further simplification usually increases bias. WHY NOT B: Less data does not solve underfitting. WHY NOT D: Changing evaluation does not fix model capacity. WHY NOT E: Random seeds do not address a systematically too-simple model.