For regression trees, the most likely goodness-of-fit measure is
_(_())_(_())
mean square-root error
◻
misclassification cost
entropy
sum of squared errors