基于HHO优化的时空水质预测模型OACSTPCD

The current situation of water resource is not optimistic,so improving the accuracy of water quality prediction models is important for water quality monitoring.In order to capture the nonlinear trend of time series data of water quality index,water quality index is mostly predicted based on neural network model.However,the existing models ignore the flow direction of the river and do not consider the influence of the water quality of the upstream monitoring points on the downstream water quality.Meanwhile,existing models mostly adjust the hyperparameters of neural networks based on the particle swarm optimization algorithm in heuristic optimization algorithms.However,the optimization algorithm still needs to set many super parameters,and improper parameter selection can easily make the model fall into local optimization.A spatio-temporal water quality prediction model(WT-CNN-LSTM-HHO)is established,Harris Hawk optimization algorithm(HHO)is used to predict downstream water quality indexes of nitrogen,phosphorus and dissolved oxygen based on upstream water quality data.The experimental results show that the proposed model can significantly improve the performance of water quality prediction,and can set fewer super-parameters and achieve higher water quality prediction accuracy.