论文部分内容阅读
分析了上海市电力月峰值负荷时间序列的特点 ,选择趋势分量和周期分量作为神经网络的输入变量 ,给出 BP神经网络辨识预测此类系统的一般方法和具体步骤。从预测的结果可以看出 ,人工神经网络辨识和预测具有趋势性及周期性波动时间序列的有效性。
The characteristics of peak load time series of Shanghai electric power month are analyzed, and the trend and periodic components are selected as input variables of neural network. The general methods and concrete steps for BP neural network identification and prediction are given. It can be seen from the prediction results that artificial neural network can recognize and predict the validity of the time series with trend and periodic fluctuation.