Adapting a Fuzzy Random Forest for Ordinal Multi-Class Classification

Fuzzy Random Forests are well-known Machine Learning ensemble methods. They combine the outputs of multiple Decision Trees to improve classification performance. Moreover, they can deal with data uncertainty and imprecision thanks use fuzzy logic. Although many tasks binary, in some situations we face problem classifying into a set ordered categories. This is particular case multi-class where order between classes relevant, for example medical diagnosis detect severity disease. In this paper, explain how binary Forest may be adapted ordinal classification. The work focused on prediction stage, not construction trees. When new instance arrives, rules activation done usual operators, but aggregation given by different trees has been redefined. particular, present procedure managing conflicting cases predicted similar support. support calculated using OWA operator that permits model concept majority agreement.