A Simple Learning System for Animal Identification

	The animal identification decision tree is represented
	as facts instead of rules – facts are easily added /
	deleted from a tree.

	Once the decision tree has reached an answer, it asks whether
	the answer is correct, and if so, then nothing more is done.
	If the answer is incorrect, however, the decision tree is
	modified to accommodate the correct answer. The answer node
	is replaced with a decision node that contains a question
	that will differentiate between the old answer that was at
	the node and the answer that was not correctly guessed.