基于随机森林算法的推移质输沙率计算研究  

Study on Bed-load Transport Rate Calculation Based on Random Forest Algorithm

作　　者：江雨润 黄尔[1] JIANG Yurun;HUANG Er(State Key Laboratory of Hydraulics and Mountain River Engineering,Sichuan University,Chengdu Sichuan 610065)

机构地区：[1]四川大学水力学与山区河流开发保护国家重点实验室,四川成都610065

出　　处：《四川水力发电》2023年第2期1-6,17,共7页Sichuan Hydropower

基　　金：第二次青藏科考,水系固体物质源-汇过程与演变(编号:2019QZKK0204)。

摘　　要：为探寻随机森林算法在预测推移质输沙率方面的效果,本文筛选了3060组天然河道输沙数据,将其划分为三种方案,得到了算法在单一流域、综合流域情况下的预测数据,并利用沙莫夫公式的计算结果作为对比。结果表明:水力参数与输沙率之间的相关性和参数在算法中的重要性并不完全一致;同一参数以不同形式出现在算法运算中时,会呈现不同的重要性;在足够的数据量支持下,随机森林算法相比于传统公式明显有更好的预测效果。In order to explore the effect of random forest algorithm in predicting the bed-load transport rate,3060 sets of bed-load transport data of natural rivers were screened and divided into three schemes to obtain the prediction data of the algorithm under the condition of single watershed and comprehensive watershed.Compared with the calculation results of Shamov formula,it shows that the correlation between hydraulic parameters and bed-load transport rate is not completely consistent with the importance of the parameters in the algorithm.When the same parameter appears in different forms in the algorithm,it will show different importance.With sufficient data,the random forest algorithm has better prediction than the traditional formula.

关 键 词：随机森林 机器学习 推移质 输沙率 

分 类 号：C32[社会学] TV142.2[水利工程—水力学及河流动力学] TV148.7