Choquet fuzzy integral-based identification

A Choquet fuzzy Integral based approach to identification of non-linear systems is investigated. The Choquet Integral replaces the maximum (minimum) operator in the information aggregation with a fuzzy integral based neuron. The identification of Choquet integral based fuzzy model is developed with strength of the rules as the input functions and unknown fuzzy densities, subject to q-measure, as the coefficients. This is a significant contribution as it leads to a class of nonadditive fuzzy systems. In addition to it, the use of q-measure provides a more flexible and powerful way of incorporating various fuzzy measures into the Integral. Simulation results show the effectiveness of the identification method.