Choquet fuzzy integral aggregation based on g-lambda fuzzy measure

It always exists the interactions between different attributes (classifiers), fuzzy integral is often chosen as an aggregation operator to describe the inherent quality which often be omitted. As we know that certain classifier maybe has different classification ability for different classes, then according to the ideas of class-indifferent fusion to obtain fuzzy densities. In this paper, g-lambda fuzzy measures and Choquet fuzzy integral are chosen to aggregate multiple outputs of trained classifiers in classification. Experimental result indicates that this methodology is effective, however the fusion accuracies are not ideal with respect to g-lambda fuzzy measures.