This paper introduces rpartScore (Galimberti, Soffritti, and Di Maso 2012), a new R package for building classification trees for ordinal responses, that can be employed whenever a set of scores is assigned to the ordered categories of the response. This package has been created to overcome some problems that produced unexpected results from the package rpartOrdinal (Archer 2010). Explanations for the causes of these unexpected results are provided. The main functionalities of rpartScore are described, and its use is illustrated through some examples.
|Journal||Journal of Statistical Software|
|Publication status||Published - Apr 2012|
- Generalized Gini impurity function
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty