Mixture of experts leveraging informer and LSTM variants for enhanced daily streamflow forecasting

Journal of Hydrology, Journal Year: 2025, Volume and Issue: unknown, P. 132737 - 132737

Published: Jan. 1, 2025

Language: Английский

---

Integrated Seasonal-Trend Decomposition Using Loess for Multi-Head Self-Attention Mechanism and Bidirectional Long Short-Term Memory Based Reference Evapotranspiration Prediction

Published: Jan. 1, 2025

Language: Английский

---

Depth prediction of urban waterlogging based on BiTCN-GRU modeling

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(4), P. e0321637 - e0321637

Published: April 23, 2025

With China’s rapid urbanization and the increasing frequency of extreme weather events, heavy rainfall-induced urban waterlogging has become a persistent pressing challenge. Accurately predicting depth is essential for disaster prevention loss mitigation. However, existing hydrological models often require extensive data have complex structures, resulting in low prediction accuracy limited generalization capabilities. To address these challenges, this paper proposes hybrid deep learning-based approach, BiTCN-GRU model, flood-prone areas. This model integrates Bidirectional Temporal Convolutional Networks (BiTCN) Gated Recurrent Units (GRU) to enhance performance. Specifically, gated recurrent units employed task. temporal convolutional network can effectively capture information features during rainfall by forward backward convolution use them as inputs GRU. Experimental results demonstrate great performance proposed achieving MAE, RMSE, R 2 values 1.56, 3.62, 88.31% Minshan Road, 3.44, 8.08, 92.64% Huaihe Road datasets, respectively. Compared such GBDT, LSTM, TCN-LSTM, exhibits higher depth. provides robust solution short-term prediction, offering valuable scientific insights theoretical support

Language: Английский

---

Decentralized multi-agent federated and reinforcement learning for smart water management and disaster response

Alexandria Engineering Journal, Journal Year: 2025, Volume and Issue: 126, P. 8 - 29

Published: April 25, 2025

Language: Английский

---

A novel fault diagnosis framework empowered by LSTM and attention: A case study on the Tennessee Eastman process

The Canadian Journal of Chemical Engineering, Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 20, 2024

Abstract In the era of Industry 4.0, substantial research has been devoted to field fault detection and diagnosis (FDD), which plays a critical role in preventive maintenance large chemical processes. However, existing studies are primarily focused on few‐shot samples process data without considering activation functions temporal diagnostic tasks. this paper, an end‐to‐end framework that combines bidirectional long short‐term memory (LSTM) with attention mechanism is proposed. preprocessing stage, special sliding time window function developed integrate multivariate containing complex information via operation such as subset extraction. Afterwards, LSTM constructed address dynamic relationship longer series observation, adopted highlight key features by assigning different weights. A case application performed enriched Tennessee Eastman (TEP), reduces bias between sample statistics larger population parameters compared studies. The metric evaluation experiments for six activations show model configured tanh can achieve optimal tradeoff tasks, providing strong benchmark subsequent research.

Language: Английский

---

Attention-based deep learning framework for urban flood damage and risk assessment with improved flood prediction and land use segmentation

International Journal of Disaster Risk Reduction, Journal Year: 2024, Volume and Issue: unknown, P. 105165 - 105165

Published: Dec. 1, 2024

Language: Английский

---