基于LSO-RF模型的阶跃型滑坡位移速率预测方法

Prediction of displacement rate of step⁃like landslide based on LSO⁃RF model

针对阶跃型滑坡在预测其位移速率时存在精度不高的问题,以泉州市安溪县尧山村阶跃型滑坡为例开展相应研究.首先,基于斯皮尔曼相关系数和灰色关联度综合分析,选取预测模型的输入特征;其次搭建结合扩展窗口法的狮群优化(LSO)-随机森林(RF)模型,提出一种适用于阶跃型滑坡位移速率预测的新方法.结果表明:综合斯皮尔曼相关系数和灰色关联度结果的特征选择方法,能弥补各自的局限性,选出最适合预测模型的输入特征组合;经过对比分析,LSO-RF模型预测阶跃型滑坡位移速率精度较高,能解决常见模型在预测阶跃型滑坡位移速率上的不足,可为阶跃型滑坡位移速率的预测提供参考.

A study was conducted to address the issue of low accuracy in predicting the displacement rate of step⁃like landslides,using the case of the step⁃like landslide in Yaoshan Village,Anxi County,Quanzhou City as an example.Firstly,the input features for the prediction model were selected based on the combined analysis of Spearman correlation coefficient and grey correlation degree.Secondly,a new method suitable for predicting the displacement rate of step⁃like landslide was proposed by the lion swarm optimization(LSO)⁃randomforests(RF)model,which incorporates the extended window method.The results show that the feature selection method,combining the Spearman correlation coeffi⁃cient and gray correlation analysis,mitigates their respective limitations and identifies the most suitable combination of input features for the prediction model.Through comparative analysis,the LSO⁃RF model demonstrates higher accuracy in predicting the displacement rate of step⁃like landslide,address⁃ing the shortcomings of common models.The conclusions can provide reference for the prediction of displacement rate of step⁃like landslide.