Research on Water Consumption Prediction Based on DWT-Informer Model
To accurately present the trend of water consumption changes and predict future water demand,a water consumption prediction method based on the DWT-Informer model is proposed.Compared with traditional methods,this prediction method has the following advantages:1)DWT decomposition of historical water consumption data can better capture the different frequency components and changing trends of water consumption signals;2)The Informer model has stronger time series modeling and prediction capabilities,which can more accurately predict future daily water consumption;3)Using a multi-headed attention mechanism to construct a global relationship between input and output is beneficial for improving parameter levels.Example analysis is conducted based on actual daily water consumption data,the results show that compared to other commonly used prediction methods,the method proposed in this paper performs excellently in MAE,RMSE,MAPE and other indicators.
water consumptionDWT decompositionmulti-headed attentionDWT-Informer model
国家科技期刊平台
NSTL
万方数据
维普
