调水调沙期小浪底水库出库泥沙组分估算研究.

In order to achieve the accurate sediment composition estimation of out reservoir in Xiaolangdi Reservoir, the estimation models for each sediment composition of out reservoir that comprehensively considering various influencing factors were established, using three different machine learning algorithms including XGBoost, KNN and GPR and based on the series data of water and sediment during the period of wa-ter-sediment regulation from 2002 to 2019. The analysis results show that it is effective to apply machine learning algorithms to estimate the sediment composition of out reservoir, and there is the good correlation between the estimated and actual value of each model. In comparison, for different sediment compositions of out reservoir, the coefficient of determination R2 of the estimation model using KNN algorithm is the highest, and the average absolute error EMAE and root mean square error ERmsEof it are all the smallest, which shows that the KNN algorithm model has higher accuracy and precision on the sediment composition estimation of out reservoir. [ABSTRACT FROM AUTHOR]