基于混沌粒子群改进支持向量机对露天矿边坡稳定性的分类预测  

Classification and Prediction of Slope Stability of Open⁃Pit Mine with Support Vector Machine Based on Chaotic Particle Swarm Optimization

作　　者：赵国彦[1] 邹景煜 王猛 ZHAO Guoyan;ZOU Jingyu;WANG Meng(School of Resources and Safety Engineering,Central South University,Changsha 410083,Hunan,China)

机构地区：[1]中南大学资源与安全工程学院,湖南长沙410083

出　　处：《矿冶工程》2024年第2期8-12,共5页Mining and Metallurgical Engineering

基　　金：国家重点研发计划项目(2018YFC0604606)。

摘　　要：为了简便有效地评估边坡稳定性状态,针对目前传统机器学习的算法选择与超参数优化等难题,提出了基于混沌粒子群优化算法的4种机器学习模型,并对其预测性能进行了对比。建立了包含221组露天矿边坡稳定性案例的数据库,其中80%的数据用于训练,20%的数据用于模型测试。4种模型预测结果及工程实例验证结果表明,基于混沌粒子群改进支持向量机模型的预测效果上总体优于其他3种机器学习模型,预测准确率88%,能够有效预测边坡稳定性,可为露天矿边坡安全提供可靠的预测结果。In order to simply and effectively evaluate slope stability,four machine learning models based on chaotic particle swarm optimization(CPSO)were proposed to solve the existing problems of algorithm selection and hyper⁃parameter optimization in traditional machine learning model,and their prediction performance were comprehensively compared among each other.A database consisting of 221 open⁃pit slope stability cases was established,in which 80%of the data were used for training and 20%for model testing.Based on the comparison between the prediction results of four models and the verification results of engineering practices,it is found that the support vector machine(SVM)based on CPSO is superior than other three machine learning models in terms of prediction of slope stability,presenting an accuracy up to 88%.Thus,it can provide a reliable prediction for the safety of slope in open⁃pit mine.

关 键 词：边坡稳定性 混沌粒子群优化 支持向量机 预测 

分 类 号：TD854[矿业工程—金属矿开采]