I'm reading the book- Intro to Data Mining by Pang-Ning Tan. Under "Bagging" it's written:
If a base classifier is stable, i.e., robust to minor perturbations in the training set, then the error of the ensemble is primarily caused by bias in the base classifier.
I understand that if base classifier is stable then it's variance will be low so bagging can't improve the variance much. But how can we conclude that the error of the ensemble is primarily caused by bias in the base classifier?
Is it because since variance is already low then the error could only be because of the bias(or noise) by Bias-Variance decomposition? Can't we directly conclude this fact about the bias without taking low variance into account?