Proceedings of the IEEK Conference (대한전자공학회:학술대회논문집)
- 2000.07b
- /
- Pages.1037-1040
- /
- 2000
The Optimal Combination of Neural Networks for Next Day Electric Peak Load Forecasting
- Konishi, Hiroyasu (Shikoku Research Institute inc) ;
- Izumida, Masanori (Ehime University) ;
- Murakami, Kenji (Ehime University)
- Published : 2000.07.01
Abstract
We introduce the forecasting method for a next day electric peak load that uses the optimal combination of two types of neural networks. First network uses learning data that are past 10days of the target day. We name the neural network Short Term Neural Network (STNN). Second network uses those of last year. We name the neural network Long Term Neural Network (LTNN). Then we get the forecasting results that are the linear combination of the forecasting results by STNN and the forecasting results by LTNN. We name the method Combination Forecasting Method (CFM). Then we discuss the optimal combination of STNN and LTNN. Using CFM of the optimal combination of STNN and LTNN, we can reduce the forecasting error.
Keywords