Data mining in load forecasting: an approach using fuzzy techniques

An alternative approach to compute the short and medium load forecasting value using a computational package using fuzzy techniques is shown. The major advantage of this system is its speed, accuracy and flexibility to find the forecasted value. The model utilizes fuzzy techniques to produce a data model for short and medium-term load forecasting studies. The engineer responsible for this activity can also introduce his/her experience before the final forecasting. A user-friendly interface has been developed to provide a man-machine interface. To build this interface, many suggestions from the operators and engineers have been incorporated. This paper also details this interface with all characteristics and features. The Previsor Program has run for more than five years at CEMIG Brazilian State Power Industry