基于OIF-Elman神经网络的燃气日负荷预测

Forecasting of Daily Gas Load Based on OIF-Elman Neural Network

与传统的 Elman 神经网络相比,采用具有输出-输入反馈机制的改进 Elman(即OIF—Elman)神经网络对燃气日负荷进行预测,不仅计入了隐层节点的反馈,而且考虑输出层节点的反馈,以便从有限的训练样本中获得更多的信息。预测结果表明,在样本较少时,无论在训练速度上,还是在预测准确度上,OIF-Elman 网络明显优于 Elman 网络。

Compared with the conventional Elman neural network, the output-input feedback （OIF） Elman neural network for forecasting of daily gas load takes into account the hidden node feedback and the output node feedback so as to obtain more information from the limited training samples. The forecasting results show that the OIF-Elman neural network is better than Elman neural network in training speed and forecasting accuracy when there are few samples.