Inductive learning is an efficient way to construct knowledge from the observation of a set of cases. It rises from the particular to the general and it provides a system with the capacity of finding by itself any useful knowledge to handle forthcoming cases. Given a set of observed cases (a so-called training set), an inductive learning algorithm is able to construct a more complex knowledge base. This paper focuses on one of the inductive learning algorithms that are most intensively used in data mining. This algorithm enables the construction of a fuzzy decision tree which represents a set of decision rules.