基于非参数GARCH的时间序列模型在日前电价预测中的应用.pdfVIP

第 36 卷 第 4 期 电 网 技 术 Vol. 36 No. 4 2012 年 4 月 Power System Technology Apr. 2012 文章编号：1000-3673 （2012 ）04-0190-07 中图分类号：TM 715 文献标志码：A 学科代码：790·625 基于非参数 GARCH 的时间序列模型 在日前电价预测中的应用 邓佳佳，黄元生，宋高峰 （华北电力大学经济与管理学院，河北省 保定市 071003 ） Day-ahead Electricity Price Forecasting Using Time Series Model Based on Nonparametric Generalized Auto Regressive Conditional Heteroskedasticity DENG Jiajia, HUANG Yuansheng, SONG Gaofeng (School of Economonics and Management, North China Electric Power University, Baoding 071003, Hebei Province, China) ABSTRACT: In electricity market the price sequence 自回归滑动平均(auto-regressive integrated moving average， fluctuates frequently, periodically and stochastically, ARIMA)模型进行预测，采用非参数 GARCH 模型对电价序 consequently price spikes often appear and it impacts the 列预测残差的随机波动率进行建模，从而提高对价格波动性 accuracy of price forecasting. Based on wavelet transform and 的预测能力和ARIMA 模型的预测精度。将该模型应用于美 nonparametric generalized auto regressive conditional 国宾夕法尼亚—新泽西— 马里兰(Pennsylvania-New Jersey- heteroscedasticity (NPGARCH) model, a time sequence model Maryland ，PJM) 电力市场的日前电价预测。算例结果表明， for the forecasting of day-ahead price was proposed. Utilizing 非参数 GARCH 模型可以更好地拟合电价序列剧烈波动的 wavelet transform, the historical price sequence was 特性，该模型能够提高电价的预测精度。 decomposed and reconstructed into a general picture sequence 关键词：电价预测；小波变换；累积式自回归滑动平均模型； and a detail sequence, and corresponding auto-regressive 非参数 GARCH 模型 integrated moving average (ARIMA) models were built respectively for price forecasting. The NPGARCH model