A general framework for estimating similarity of datasets and decision trees: exploring semantic similarity of decision trees