A load curve based fuzzy modeling technique for short-term load forecasting (Q1874048)

The authors propose a fuzzy rule-based model for short-term forecasting. Its development is carried out in a general setting of Computational Intelligence and comprises three main phases; that is (a) determination of the structure of the model, (b) selecting input variables through genetic optimization (genetic algorithm), and (c) fine tuning of the model through a hybrid of the genetic/least square error algorithm. The effectiveness of the model is augmented by the realization of the conclusion part in the form of cubic B-splines. The study comes with an extensive experimental part including some comparative analysis exploiting other neural network models.