基于TCN-LSTM-QR的地表水水质预测模型

Surface Water Quality Prediction Model Based on TCN-LSTM-QR

随着时间推移,地表水的营养元素含量不断增加,这已成为一个严重的水环境挑战。本文提出了一种新的方法来监测国家地表水水质,该方法基于时间卷积神经网络和长短期记忆神经网络,可以更好地分析历史监测数据,并将总磷和总氮作为导致富营养化的重要指标。通过LSTM和分位数回归技术,我们可以有效地预测地表水的水质。实验结果表明,所提出的模型不仅有高精度的点预测结果,还可获得某一置信水平的区间预测结果。

Over time,the nutrient content of surface water has been increasing,which has become a serious water environment challenge already.This paper proposes a method based on Temporal Convolutional Neural Network(TCN),Long Short-Term Memory Neural Network(LSTM)and Quantile Regression(QR)to measure the historical monitoring data of the national surface water quality automatic monitoring real-time data distribution system,with total phosphorus and total nitrogen as a significant indicator of eutrophication.Memory Neural Network(LSTM)and Quantile Regression(QR)for surface water quality prediction.The proposed model's experimental results demonstrate not only its high accuracy in point prediction,but also its ability to generate interval prediction results with a certain degree of assurance.