Discretization oriented to Decision Rules Generation

摘要: Many of the supervised learning algorithms only work with spaces dis- crete attributes. Some methods proposed in bibliography focus on cretization towards generation decision rules. This provides a new algorithm called USD (Unparametrized Supervised Discretization), which transforms infinite space values continuous attributes finite group intervals purpose using these rules, such way that rules do not loose accuracy or goodness. Stands out fact that, contrary to other methods, doesn't need parameterization.