基于PCA-GA-SVM模型的地表下沉系数预测  

Prediction of Surface Subsidence Coefficient Based on PCA-GA-SVM Model

作　　者：史耀凡 栾元重[1] 于水[1] 郭凯维 SHI Yaofan;LUAN Yuanzhong;YU Shui;GUO Kaiwei(School of Surveying,Mapping and Spatial Information,Shandong University of Science and Technology,Qingdao,Shandong 266590,China)

机构地区：[1]山东科技大学测绘与空间信息学院,山东青岛市266590

出　　处：《矿业研究与开发》2022年第2期65-69,共5页Mining Research and Development

基　　金：山东省自然科学基金项目(ZR2020MD024).

摘　　要：为解决地表下沉系数难以精准预测的问题,提出了基于主成分分析(PCA)的遗传算法(GA)优化的支持向量机(SVM)模型。利用主成分分析法对45个矿区的7个地表移动参数影响因子进行降维,提取出新的主成分,同时通过遗传算法获取能够优化支持向量机的最优惩罚参数和最优核函数参数,将主成分输入优化后的支持向量机,对比地表下沉系数的预测值与真实值之间的差距,计算平均相对误差,并与SVM,PCA-SVM,GA-SVM 3种模型的平均相对误差对比。结果表明:PCA-GA-SVM模型的平均相对误差可以达到7.01%,精度更高,更能准确地预测地表下沉系数。In order to solve the problem that the surface subsidence coefficient is difficult to accurately predict,support vector machine(SVM)model optimized by genetic algorithm(GA)based on principal component analysis(PCA)was proposed.The principal component analysis method was used to reduce the dimension of 7 surface movement parameter influencing factors in 45 mining areas,and the new principal components were extracted.Meanwhile,the optimal penalty factor and the optimal kernel function parameters that could optimize the support vector machine were obtained through the genetic algorithm.The support vector machine optimized by inputting the principal components was used to compare the difference between the predicted value and the true value of surface subsidence coefficient.The average relative error was compared with that of SVM,PCA-SVM,and GA-SVM models.The results show that the average relative error of PCA-GA-SVM model can reach 7.01%,with a higher accuracy»which can accurately predict the surface subsidence coefficient.

关 键 词：地表下沉系数 主成分分析法 遗传算法 支持向量机 

分 类 号：TD854^(+).6[矿业工程—金属矿开采]