基于TCK-LSTM-ATT模型的城市用水量预测

Prediction of Urban Water Consumption Based on TCK-LSTM-ATT Model

针对LSTM神经网络模型的计算量较大,不可控的自主选择过程以及容易过拟合等问题,提出了TCK-LSTM-ATT模型,利用卷积核对数据特征进行提取合并,采用注意力机制对重要数据进行加权的组合模型方法。为了验证该模型对于供水量预测的准确性,利用中国东北某市2019年到2020年的某供水管网系统供水数据进行验证。实验结果表明,与普通模型相比,组合模型的预测误差减少约20%,R^(2)值约为9.5,取得了较好的预测效果。

Aiming at the problems of large amount of calculation,uncontrollable independent selection process and easy over fitting of LSTM,a combined model method is proposed.The model uses convolution to extract features and merge features,uses attention mechanism to weight important data.The model is called TCK-LSTM-ATT model.In order to verify the accuracy of this model in water supply prediction,the water supply data of a urban water supply network system in Northeast China from 2019 to 2020 are used for verification.Compared with the ordinary model,the prediction error of this model is reduced by about 20%,and the R^(2) value is about 9.5.It is considered that a good prediction effect is achieved in the experiment.