An Integrated Framework of GRU Based on Improved Whale Optimization Algorithm for Flood Prediction

Abstract

Accurate prediction of floods is the first step in formulating flood control strategies and reducing flood disasters. This research proposes a deep learning model based on Gate Recurrent Unit (GRU), Random Forest Algorithm (RF), Whale Optimization Algorithm (WOA) and Optimal Variational Mode Decomposition (OVMD) for flood prediction. First, the random historical time series is decomposed using OVMD. Secondly, combined with the RF feature importance measurement, select features with high importance to obtain the optimal input set. Third, use the GRU model to predict all sub-models, and use the WOA algorithm to optimize the hyperparameters in the GRU model. This study also proposes a hybrid strategy to improve the traditional WOA algorithm and enhance the optimization ability of the WOA algorithm. Finally, the prediction results of all sub-modes were aggregated to generate the final prediction result. The model was validated using data from three hydrological stations in the upper, middle and lower reaches of the Minjiang river basin in China. Through the results of the experiment, it can be seen that the proposed prediction model can effectively predict the flood time series, and has better accuracy than other models.

A novel hybrid model for hourly streamflow and water level prediction from radar reflectivity using deep learning approaches

“…The history information is updated by controlling these three gates. The forget gate controls the information retained to the current last moment; the input gate is responsible for regulating the current input; the output gate is responsible for regulating the current output (Ji et al, 2021). The function of the GRU is similar to the LSTM, but the GRU's structure is simpler than that of the LSTM.…”

“…The function of the GRU is similar to the LSTM, but the GRU's structure is simpler than that of the LSTM. The GRU merges the forget and input gates of the LSTM into one update gate, and the output gate becomes one reset gate (Ji et al, 2021). Through the structure of GRU, the input information from the previous module is added, and then the new state of output in the current module is obtained through calculation (Han et al, 2021).…”

“…Through the structure of GRU, the input information from the previous module is added, and then the new state of output in the current module is obtained through calculation (Han et al, 2021). The weight values in these structures are continuously adjusted through learning until the output meets the requirements, and finally achieve the purpose of screening information, controlling data flow, determining whether the information is transmitted, and simulating the reasoning process (Ji et al, 2021). This helps the GRU model address the problem of the vanishing gradient of recurrent neural networks (RNNs) and reduce the training time of the model (Zhang & Wang, 2022).…”

A Systematic Review of Deep Learning Applications in Streamflow Data Augmentation and Forecasting

“…combined PCA with GRUs and LSTMs separately for the Muskegon River and the Pearl River, China, and showed that integrating meteorological data into the training increases the accuracy Yuan et al (2021). andXiao and Wang (2021) employed EEMD with LSTMs Ji et al (2021). employed GRUs with Random Forest, Whale Optimization Algorithm (WOA) and Optimal Variational Mode Decomposition (OVMD) for daily streamflow of the Minjiang river basin in China.Among comparative studies,Guo et al (2021) trained LSTMs, GRUs, and SVMs for 25 reservoirs in China and concluded that even though LSTMs and GRUs performed similarly, GRUs trained faster.…”