Fuzzy hidden Markov predictor in electric load forecasting

We present a new hybrid system that merges fuzzy logic with dynamic Bayesian networks (DBN's): the fuzzy hidden Markov predictor. It is a modification of the hidden Markov model, a particular case of DBN's, in order to enable it to predict continuous values of a time series. A DBN is a Bayesian network that represents a temporal probability model. This hybrid system is applied to the task of monthly electric load single-step forecasting and successfully compared with three regression-by-discretization systems, two fuzzy hybrid systems, two Kalman filter models, and Box-Jenkins and Winters exponential smoothing. The employed time series present a sudden significant changing behavior at their last years, as it occurs in an energy rationing.