In this paper, a new Fuzzy Inference Net (FIN)-based method is proposed for electricity price forecasting. FIN makes use of Fuzzy Self-organization Mapping (FSOM) to evaluate the attribute probability of unknown data for the predetermined clusters. This paper focuses on the predicted zone rather than the predicted value due to the difficulty of the price forecasting with high volatility. After the zone of the electricity price is sectionalized into several zones, this paper constructs FIN at each zone that evaluates the attribute probability of unknown input data to each zone. The proposed method is successfully applied to the one-day-ahead electricity price of real data.
展开▼
