● A novel deep learning framework for short-term water demand forecasting.● Model prediction accuracy outperforms other traditional deep learning models.● Wavelet multi-resolution analysis automatically extracts key water demand features.● An analysis is performed to explain the improved mechanism of the proposed method. Short-term water demand forecasting provides guidance on real-time water allocation in the water supply network, which help water utilities reduce energy cost and avoid potential accidents. Although a variety of methods have been proposed to improve forecast accuracy, it is still difficult for statistical models to learn the periodic patterns due to the chaotic nature of the water demand data with high temporal resolution. To overcome this issue from the perspective of improving data predictability, we proposed a hybrid Wavelet-CNN-LSTM model, that combines time-frequency decomposition characteristics of Wavelet Multi-Resolution Analysis (MRA) and implement it into an advanced deep learning model, CNN-LSTM. Four models - ANN, Conv1D, LSTM, GRUN - are used to compare with Wavelet-CNN-LSTM, and the results show that Wavelet-CNN-LSTM outperforms the other models both in single-step and multi-steps prediction. Besides, further mechanistic analysis revealed that MRA produce significant effect on improving model accuracy. 相似文献
● This study explored the long-term association by double robust additive models. ● Individual exposure concentrations were assessed by integrating GAM, LUR and BPNN. ● PM2.5, SO2 and NO2 are positively associated with cerebrovascular disease. ● CO could reduce the risk of cerebrovascular disease with the highest robustness. ● The elderly, women and people with normal BMI are at higher risk for air pollution. The relationship between air pollution and cerebrovascular disease has become a popular topic, yet research findings are highly heterogeneous. This study aims to investigate this association based on detailed individual health data and a precise evaluation of their exposure levels. The integrated models of generalized additive model, land use regression model and back propagation neural network were used to evaluate the exposure concentrations. And doubly robust additive model was conducted to explore the association between cerebrovascular disease and air pollution after adjusted for demographic characteristics, physical examination, disease information, geographic and socioeconomic status. A total of 25097 subjects were included in the Beijing Health Management Cohort from 2013 to 2018. With a 1 μg/m3 increase in the concentrations of PM2.5, SO2 and NO2, the incidence risk of cerebrovascular disease increased by 1.02 (95% CI: 1.008–1.034), 1.06 (95% CI: 1.034–1.095) and 1.02 (95% CI: 1.010–1.029) respectively. Whereas CO exposure could decrease the risk, with an odds ratio of 0.38 (95% CI: 0.212–0.626). In the subgroup analysis, individuals under the age of 50 with normal BMI were at higher risk caused by PM2.5, and SO2 was considered more hazardous to women. Meanwhile, the protective effect of CO on women and those with normal BMI was stronger. Successful reduction of long-term exposure to PM2.5, SO2 and NO2 would lead to substantial benefits for decrease the risk of cerebrovascular disease especially for the health of the susceptible individuals. 相似文献