说明：  Elman神经网络建立建筑物电力负荷预测模型中遇到的几个关键问题有，数据归一化处理、输入输出样本的选取、隐含层节点数的确定；分别建立Elman神经网络模型，并利用某栋建筑物实际历史电力负载数据进行预测，分析比较与实际数据值的预测精度，得出了一个有效的数据预测模型。
(Several key problems encountered in building power load forecasting model based on Elman neural network are data normalization, selection of input and output samples, and determination of the number of hidden layer nodes. Elman neural network models are established respectively, and the actual historical power load data of a building are used for forecasting. Effective data prediction model.)