An experimental study on 45 ordinal datasets finds Ordinal Gini the strongest splitting criterion among those tested, lowering MAE by over 3.02% versus nominal Gini.
Piccarreta, Classification trees for ordinal variables, Computational Statistics 23 (3) (2008) 407–427
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Splitting criteria for ordinal decision trees: an experimental study
An experimental study on 45 ordinal datasets finds Ordinal Gini the strongest splitting criterion among those tested, lowering MAE by over 3.02% versus nominal Gini.