Averaging operators in fuzzy classification systems

The averaging operators used for aggregation of outputs of fuzzy classification rules are examined. Several methods used in fuzzy rule-based classification systems are discussed and related multi-polar averaging operators are shown. Further, new aggregation methods based on the idea used for the definition of OWA operators are described. The connection between multi-polar averaging operators and the multi-polar Choquet integral is shown and using this connection the conditions under which are the respective averaging operators monotone are studied. Several examples of special multi-polar OWA operators with a relation to existing bipolar aggregation operators are included.