Precipitation Nowcasting in Dar es Salaam: Comparative Analysis of LSTM and Bidirectional LSTM for Enhancing Early Warning Systems

Journal of Geoscience and Environment Protection, Journal Year: 2025, Volume and Issue: 13(04), P. 327 - 342

Published: Jan. 1, 2025

Language: Английский

---

High-performance traffic volume prediction: An evaluation of RNN, GRU, and CNN for accuracy and computational trade-offs

BIO Web of Conferences, Journal Year: 2024, Volume and Issue: 148, P. 02034 - 02034

Published: Jan. 1, 2024

Predicting urban traffic volume presents significant challenges due to complex temporal dependencies and fluctuations driven by environmental situational factors. This study addresses these evaluating the effectiveness of three deep learning architectures— Recurrent Neural Network (RNN), Gated Unit (GRU), Convolutional (CNN)—in forecasting hourly on Interstate 94. Using a standardized dataset, each model was assessed predictive accuracy, computational efficiency, suitability for real-time applications, with Mean Absolute Percentage Error (MAPE), Root Square (RMSE), R 2 coefficient, computation time as performance metrics. The GRU demonstrated highest achieving MAPE 2.105%, RMSE 0.198, 0.469, but incurred longest 7917 seconds. Conversely, CNN achieved fastest at 853 seconds, moderate accuracy (MAPE 2.492%, 0.214, 0.384), indicating its real- deployment. RNN exhibited intermediate performance, 2.654% 0.215, reflecting limitations in capturing long-term dependencies. These findings highlight crucial trade- offs between underscoring need selection aligned specific application requirements. Future work will explore hybrid architectures optimization strategies enhance further feasibility management.

Language: Английский

---

Precipitation Prediction Using Long Short-Term Memory Networks: Improving Seasonal Rainfall Forecast Accuracy for Flood and Drought Prevention

Published: Aug. 16, 2024

Language: Английский

---