Comparative Study of Short-Term Electric Load Forecasting

In this paper, we performed short-term electric load forecasting using three methods and compared each results. We classified before making a forecasting model using K-means and k-NN to eliminate error from calendar based classification. Classified load data used as inputs of forecasting model. We compared three methods such as ANN, SES, GMDH. We carried out 1-day ahead prediction for two weeks, January 10 to 16, March 14 to 20, 2011 using hourly Korean electric load data. The results of forecasting, all methods were mostly good in general without applying meteorological data. Most of them, GMDH expressed the most performance in MAPE except for Saturday.