Predicting Water Quality Based On EEMD And LSTM Networks

Water is crucial for all types of life. The security of water quality genuinely influences human wellbeing, fishery economy and agrarian. Considering the characteristics of water quality are dynamic, nonlinear and complex, a novel water quality prediction combined model based on Ensemble Empirical Mode Decomposition (EEMD) and Long Short-Term Memory (LSTM) network is proposed in this paper. In the pactical work, the proposed model was trained by using the Yangtze River water quality history data which was measured daily from September 2009 to November 2017. The dataset of water quality was pretreated by ensemble empirical mode decomposition. For several imf subsequences, a LSTM submodel is established for each subsequence, and the prediction result of each submodel was added to compose the final prediction result. Finally, to obatain the illustration and verification, the method is compared with other traditional network models on practical experiments. The experimental results demonstrated that the proposed model has the best performance.