Classification Trees with Imprecise Probabilities
Description
Creation of imprecise classification trees. They rely on
probability estimation within each node by means of either the
imprecise Dirichlet model or the nonparametric predictive
inference approach. The splitting variable is selected by the
strategy presented in Fink and Crossman (2013)
,
but also the original imprecise information gain of Abellan and
Moral (2003) is covered.