Case-based reasoning for short term load forecasting

The case-based reasoning (CBR) is introduced into short term load forecasting. The bottle-neck difficulty in knowledge acquaintance of rule-based expert system is overcome by using another knowledge resource-successful forecasting cases in the past. The forecasting processes in the past are retained as cases in the case-base. New forecasting problems are solved by using the conclusions of previous similar cases retrieved in the case-base. The practical instances show that for the short term load forecasting problem the case-based reasoning method outperforms a single mathematic model on the forecasting accuracy. Therefore, it has wide applicability.