オープンデータに基づく電力スポット市場のエリア価格予測^——複数予測スキームの統合アプローチの検討^——

Forecasting the electricity spot market price is important in planning efficient power generation and demand schedules. It has become common for electric power utilities to use several commercial products of market price forecasts in their daily operations. Therefore, this study proposes a method of combining forecasts to use these products efficiently and improve forecasting accuracy. The method aims to capture the features of the component models’ results considering situations by using open data as explanatory variables. The result estimated as the less prediction error in some situations obtains high confidence during combining forecasts in same situation. To evaluate the effectiveness of the method of combining forecasts, numerical experiments were conducted on all kinds of area price in the Japan Electric Power Exchange. To the best of our knowledge, it is the first paper that conducted forecasts for all nine area prices and reported the forecasting accuracy of the proposed forecasting method. The results indicate that the method of combining forecasts is a promising way to improve the accuracy of each component forecasts.