基于ARIMA-GARCH模型的现货电价预测  

Spot Electricity Price Forecasting Based on ARIMA-GARCH Model

作　　者：刘琰[1] 刑薇[1] 丁乐群[1] 徐越[1] 韩强[1] 王宇拓[1] 

机构地区：[1]东北电力大学经济管理学院,吉林省吉林市132012

出　　处：《能源技术经济》2012年第2期59-63,共5页Electric Power Technologic Economics

摘　　要：通过建立ARIMA预测模型对现货电价进行预测,并对ARIMA模型存在的异方差问题通过GAR CH模型进行修正。实证算例中,采用北欧四国电力市场数据,与AR IMA和灰色GM(1,1)模型进行比较,表明ARIMA-GARCH模型的预测精度更高,预测误差更小。The ARIMA model is established to predict the spot electricity price, and the heteroscedasticity that exists in the ARIMA model is corrected with the GARCH model. In the empirical example, data from the Nordic electricity market are adopted, and comparisons are conducted between the ARIMA model and the Gery GM （1,1） model. The results suggest that the AR1MA-GARCH model has higher accuracy and smaller prediction errors.

关 键 词：电价 模型 时间序列 预测 

分 类 号：TK01[动力工程及工程热物理] TM715[电气工程—电力系统及自动化]