Multiple Function Approximation - A New Approach Using Complex Fuzzy Inference System

A complex-fuzzy machine learning approach to function approximation for multiple functions is proposed in this paper. The proposed approach involves the utility of complex-valued vector outputs by a novel complex-fuzzy model using complex fuzzy sets and the famous PSO-RLSE hybrid algorithm for machine learning of the model. An experiment was used to test the proposed approach for the ability of approximating four functions simultaneously. With the experimental result, the performance by the proposed model is promising and the proposed approach is compared to other methods. With complex fuzzy sets, the proposed approach has shown the excellent capability of function approximation for multiple functions using one single model with good performance.