Percentile combining classifier
W = PERC(V,P)
If V = [V1,V2,V3, ... ] is a set of classifiers trained on the same classes and W is the percentile combiner: it selects the class defined by the percentile of the outputs of the input classifiers. This might also be used as A*[V1,V2,V3]*PERC(,P) in which A is a dataset to be classified.
If it is desired to operate on posterior probabilities then the input classifiers should be extended like V = V*CLASSC;
The base classifiers may be combined in a stacked way (operating in the same feature space by V = [V1,V2,V3, ... ] or in a parallel way (operating in different feature spaces) by V = [V1;V2;V3; ... ]