The goal of inductive learning systems is to generate rules for later use. Application of rules may present some problems, however. We are concerned with the problem of several rules covering an example to be classified. We need a way for deciding which rule is to be followed. This is often referred as the conflict resolution strategy. Alternatively, we may decide to combine different opinions. Several methodologies exist to solve these problems. In the following sections three different strategies are presented. Each strategy attempts to deal with the problem of uncertainty caused, for instance, by unknown attribute values or incomplete description of examples.