基于小波神经网络的电价滚动预测模型

探讨了基于小波分析和神经网络的3种短期电价预测模型,比较了提前1步的滚动预测与提前N步的预测方法.采用预测误差概率分布作为预测误差的评价指标,并以美国加州电力市场的实际运行数据为基础,连续预测该市场1个月的电价.结果表明:提出的具有滚动预测概念的模型III具有良好的预测精度,其误差分布还显示出该模型具有较高的预测置信度.

Electricity Price Forecasting Model Based on Wavelet-Neural Network

Three short-term forecasting models based on wavelet analysis and neural network are studied.The methods of one step ahead forecasting and N steps ahead forecating are compared.They are also analyzed using errors'probability distribution.To compare the results,the research forcasts one month continuously using three models based on the historical hourly electricity prices of the California electricity market.The results show that the forecasting model III is good to improve the accuracy of prediction,The error probability distribution of forecasting model III also indicates that it has a good confidence degree.