摘要
在使用累积式自回归移动平均法(ARIMA)进行中期电力负荷预测时,所得残差序列具有明显规律。电力负荷数据可使用线性和非线性成分叠加表示,为弥补传统ARIMA时间序列预测法忽略非线性的缺陷,引入LIBSVM支持向量机挖掘数据残差非线性规律,并将LIBSVM预测残差与ARIMA预测结果相叠加,达到更高的精度。使用ARIMA-LIBSVM组合模型进行实例预测,结果表明:该模型能够提高预测精度。
The actual prediction shows that when the cumulative autoregressive integrated moving average(ARIMA) method is used for medium-term power load forecasting,the obtained residual sequence has obvious regularity.Power load data has linear and nonlinear components.In order to make up for the defects that traditional ARIMA time series prediction method ignores nonlinear,LIBSVM support vector machine is introduced to mine the nonlinear law of data residual,and the LIBSVM prediction residual is superimposed with ARIMA prediction results to achieve higher accuracy.The model is used to predict the example,and the results show that ARIMA-LIBSVM model can improve the prediction accuracy.
作者
李晨
尹常永
李奇洁
李春雷
LI Chen;YIN Chang-yong;LI Qi-jie;LI Chun-lei(School of Electric Power,Shenyang Institute of Engineering,Shenyang 110136,Liaoning Province)
出处
《沈阳工程学院学报(自然科学版)》
2023年第1期49-55,共7页
Journal of Shenyang Institute of Engineering:Natural Science
关键词
负荷预测
时间序列分析
支持向量机
残差修正
Load forecasting
Time series analysis
Support vector machine
Residual correction