Extração de regras de conhecimento a partir de redes neurais artificiais aplicadas para a previsão de demanda por energia elétrica

Abstract

This work presents a methodology for the extraction of rules from Artificial Neural Networks (ANN) trained to forecast the electric load demand. The rules have the ability to express the knowledge regarding the behavior of load demand acquired by the network during training process. The rules are presented to the user in an easy to read format, such as

IF premise THEN consequence. Where premise relates to the input data submitted to the network, and consequence appears as a linear equation describing the output to be presented by the network, should the premise part holds true. The rule extraction technique presented here, combined with the predictive abilities of the neural networks, constitutes an important tool for tactical and strategic decision making within the electricity supply companies.