Training a FIS with EPSO under an Entropy Criterion for Wind Power prediction

This paper summarizes efforts in understanding the possible application of information theoretic learning principles to power systems. It presents the application of Renyi's entropy combined with Parzen windows as a measure of information content of the error distribution in model parameter estimation in supervised learning. It illustrates the concept with an application to the prediction of power generated in a wind park, made by Takagi-Sugeno fuzzy inference systems, whose parameters are discovered with an EPSO-evolutionary particle swarm optimization algorithm