摘要
针对常规支持向量机预测模型在变形数据处理预测中的不足,本文提出了一种基于改进灰狼算法的支持向量回归模型。重新定义了灰狼算法中的收敛因子,并引入多项式变异算子,使得算法在收敛方面得到改善;将具有局部特征的柯西核函数和具有全局特征的多项式核函数进行组合,以此来综合核函数的两种不同特性,提高预测数据集的整体精度。采用基坑监测项目数据对模型预测能力进行实验,并与其他模型进行对比分析。结果表明,本文模型对结构变形发展演化的非线性特征拟合精度更高,可以应用到时间序列变化的数据预测处理。
Aiming at the insufficiency of support vector machine prediction model in deformation data processing and prediction,this paper proposed a support vector regression model based on improved gray wolf algorithm.The convergence factor in the gray wolf algorithm was redefines and a polynomial mutation operator was introduced to improve the algorithm's convergence.The Cauchy kernel function with local characteristics and the polynomial kernel function with global characteristics were combined to synthesize the two different characteristics of the kernel function and improve the overall accuracy of the prediction data set.The model’s predictive ability was tested with the data of foundation pit monitoring projects,and other models were used for comparative experiments.The results showed that the model in this paper had higher fitting accuracy and better performance for the nonlinear characteristics of structural deformation evolution.The research results can be applied to data prediction of time series changes.
作者
徐培云
周兆玺
闵阳
袁辉
陈玉奇
XU Peiyun;ZHOU Zhaoxi;MIN Yang;YUAN Hui;CHEN Yuqi(China Railway Siyuan Survey and Design Group Company Limited, Wuhan Hubei 430063, China;School of Geography and Information Engineering, China University of Geosciences, Wuhan Hubei 430078, China;Wuhan Geotechnical Engineering and Surveying Company Limited,Wuhan Hubei 430022,China)
出处
《北京测绘》
2022年第5期537-542,共6页
Beijing Surveying and Mapping
基金
中铁第四勘察设计院集团有限公司科研课题(2020D089)。
关键词
变形监测
支持向量机
核函数
收敛因子
灰狼优化算法
deformation monitoring
support vector machines(SUM)
kernel function
convergence factor
Grey Wolf optimization algorithm(GWO)
