河北地质大学学报2025，Vol.48Issue(1)：69-77,9.DOI:10.13937/j.cnki.hbdzdxxb.2025.01.008
基于特征选取与SVR优化的软土路基弹性模量预测模型
Prediction Model of Elastic Modulus of Soft Soil Subgrade Based on Feature Selection and SVR Optimization
摘要
Abstract
Based on the prediction method of the parameters and the support vector regression(SVR)model,the elastic modulus.A modified particle swarm algorithm(PSO)is designed for the long and long intensive search to optimize the parameters of the elastic modulus prediction model of support vector regression to find the global optimal solution.The study shows that according to the correlation coefficient,mean squared error and absolute error,compared with other models,PSO-SVR model is easier to jump out of the local optimal solution,and has high accuracy,stability and robustness.关键词
特征选取/PSO/SVR/弹性模量Key words
feature selection/PSO/SVR/elastic modulus分类
地质学引用本文复制引用
肖乾,袁颖..基于特征选取与SVR优化的软土路基弹性模量预测模型[J].河北地质大学学报,2025,48(1):69-77,9.基金项目
安徽省高等学校自然科学项目(2021ZK02) （2021ZK02）
安徽省高等学校哲学社会科学项目(2022AH052659) （2022AH052659）
