摘要
边坡稳定性预测对工程安全及地质灾害防治极其重要,目前机器学习在边坡稳定性预测较广泛,例如BP神经网络、支持向量机(SVM)、极限学习机(ELM)等。但传统的ELM模型在预测边坡稳定性时存在易陷入局部最小值、难以选择合适学习率的问题,针对此问题,提出了一种基于主成分分析法(PCA)和爬行动物搜索法(RSA)并行优化极限学习机(ELM)的边坡稳定性预测模型。此模型利用PCA算法对数据进行降维,减少数据的冗余性,并利用RSA算法优化ELM模型的输入层权值和隐含层偏置,极大地提高了模型的预测精度和预测效率。将传统的ELM模型、RSA-ELM模型、PCA-SVM模型及PCA-RSA-ELM 4种模型进行对比,从而得到PCA-RSA-ELM模型在边坡稳定性预测这类问题上的精确性更高,为边坡稳定性预测分析提供新的思路,对防灾减灾及保护国民经济安全具有重大意义。
Slope stability prediction and analysis are very important for engineering safety and geological disaster prevention.At present,machine learning is widely used in slope stability prediction,such as BP neural network,support vector machine(SVM),extreme learning machine(ELM)and so on.However,the traditional ELM model is prone to fall into the local minimum value and is difficult to select the appropriate learning rate when predicting slope stability.Aiming at this problem,a slope stability prediction model based on principal component analysis(PCA)and reptile search method(RSA)parallel optimization limit learning machine(ELM)is proposed in this paper.This model uses the PCA algorithm to reduce the dimension of data and reduce the redundancy of data,and uses the RSA algorithm to optimize the input layer weight and hidden layer bias of ELM model,which greatly improves the prediction accuracy and efficiency of the model.By comparing the traditional ELM model,RSA-ELM model,PCA-SVM model and PCA-RSA-ELM model,it′s found that the PCA-RSA-ELM model has higher accuracy in slope stability prediction,which provides a new idea for slope stability prediction analysis.In the meantime,it is of great significance to disaster prevention and reduction and protection of national economic security.
作者
陈家豪
张燕
杜明芳
黄海荣
徐志军
陈旭
CHEN Jiahao;ZHANG Yan;DU Mingfang;HUANG Hairong;XU Zhijun;CHEN Xu(College of Civil Engineering,Henan University of Technology,Zhengzhou 450001,China;Henan Key Labratory of Grain Storage Facility and Safety,Zhengzhou 450001,China)
出处
《金属矿山》
CAS
北大核心
2024年第6期191-198,共8页
Metal Mine
基金
国家自然科学基金面上项目(编号:51978247)
河南工业大学青年骨干教师培育计划(编号:21420155)
河南省粮油仓储建筑与安全重点实验室开放课题(编号:2020KF-B01)。
关键词
安全工程
边坡稳定性
极限学习机
PCA
降维
爬行动物搜索
混淆矩阵
safety engineering
slope stability
extreme learning machine
PCA dimension reduction algorithm
reptile search algorithm
confusion matrix
