摘要
分析了当前河流纵向离散系数的各种计算方法及其缺陷,提出了一种新的计算河流纵向离散系数的方法。该方法基于SVM理论建立SVR回归模型,将河流的平均水深H、宽度W、流速v以及剪切流速v*作为模型的输入变量,将河流的纵向离散系数EX作为模型的输出变量,选取22组数据用于模型的训练和验证。预测的平均相对误差为0.29%,结果表明,这种方法可以简单有效地预测河流纵向离散系数。
The various calculation methods and defects of determining river longitudinal dispersion coefficient were analyzed,a nw method for calculating the river longitudinal dispersion coefficient was put forward. The method was based on SVM theory to establish a SVR regression model. The input variables were mean river depth H,width W,velocity v and friction velocity v＊. The output variable was longitudinal dispersion coefficient EX. 22 groups of data were selected to train and test model,among which,20 groups were selected as training data and the rest groups were used as validation data. The mean relative error of the model prediction was 0.29.The result indicated that the SVM algorithm can predict river longitudinal dispersion coefficient efficiently and simply.
出处
《安徽农业科学》
CAS
北大核心
2010年第23期12630-12631,12670,共3页
Journal of Anhui Agricultural Sciences
关键词
河流纵向离散系数
SVR
平均相对误差
River longitudinal dispersion coefficient
Support vector regression
Mean relative error
