|
A new forecast engine based on data fusion for electricity price forecasting
|
Ali Darudi1   , Masoud Bashari2 , Mohammad Hossein Javidi *2 |
1- Ferdowsi University of Mashhad
2- Mashhad |
|
|
Abstract: (10286 Views) |
| In restructured electricity markets, accurate price forecasting plays an important role for all market participants. Due to the complexity and distinct nature of the electricity price, a single forecast engine cannot capture and model all different patterns in price signals. As a result, to improve forecast accuracies, this paper proposes a hybrid method to use advantages of several forecast engines simultaneously. In the proposed method, three primary engines, artificial neural networks (ANN), adaptive neuro-fuzzy inference systems (ANFIS), and autoregressive moving average (ARMA), provides three independent forecasts of the price. Then, a new fusion algorithm combines these three forecasts to obtain a unified single price forecast. The proposed method obtains feedback from previous error of the primary forecast engines to adjust their effect on the final forecast. The proposed method is evaluated using price data of Spanish electricity market. Results indicate that the proposed method outperform each primary forecasting engine. |
|
| Keywords: Electricity market, price forecasting, data fusion, ordered weighted average (OWA), Fuzzy Choquet integral |
|
|
Full-Text [PDF 771 kb]
(1786 Downloads)
|
Type of Study: Research |
Subject:
Special
Received: 2014/10/6 | Accepted: 2015/03/16 | Published: 2015/04/15
|
|
|
|
|
|
