Flow-based grey single-layer perceptron with fuzzy integral

A flow-based perceptron constructed on the basis of the nonadditive grey single-layer perceptron (GSLP) is proposed. In common with the nonadditive GSLP, the proposed perceptron has the property that it measures the grades of relationship between input patterns and a typical pattern using gray relational analysis (GRA) and a Choquet fuzzy-integral-based neuron. However, the proposed perceptron further uses the single criterion net flow for each criterion, instead of the original performance value, to perform GRA. All flows representing preference information among patterns can be generated using the preference relation to gauge the intensity of preference for one pattern over another on each criterion. Experimental results further demonstrate that the generalization ability of the proposed perceptron performs well compared to that of the nonadditive GSLP.