基于CRITIC赋权法与PSO-SVR模型的滑坡地表位移预测研究
DOI:
作者:
作者单位:

1.东华理工大学江西省放射性地学大数据技术工程实验室;2.东华理工大学地球物理与测控技术学院

作者简介:

通讯作者:

中图分类号:

P642.22

基金项目:

江西省放射性地学大数据技术工程实验室(JELRGBDT202206); 江西省自然科学(20212BAB203004)。


Research on Landslide Surface Displacement Prediction Based on CRITIC Weight Method and PSO-SVR Model
Author:
Affiliation:

1.Jiangxi Engineering Laboratory of Radiofrequency Big Data Technology,East China University of Technology;2.School of Geophysics and Measurement and Control Technology,East China University of Technology,Nanchang

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    针对支持向量机在滑坡位移预测中输入项权值无差异,从而影响模型预测精度的问题,提出一种基于CRITIC(Criteria Importance Though Intercrieria Correlation)赋权法与粒子群算法(particle swarm optimization, PSO)优化支持向量回归机(support vector regression, SVR)的滑坡位移预测模型。该模型首先采用皮尔逊相关性分析法,选取与模型输出项相关性较强的3项影响因素,然后由CRITIC赋权法求得对应权值,将加权后的训练集输入基于CRITIC赋权法与PSO-SVR的预测模型,以实现对滑坡地表位移的预测。结果表明:相比SVR、PSO-SVR以及基于熵权法与PSO-SVR的预测模型,本模型具有良好的泛化能力,均方根误差和判定系数分别比未赋权模型降低38.24%和提高6.64%,能有效提高预测精度,预测效果优于其他对比模型。

    Abstract:

    A landslide displacement prediction model is proposed based on CRITIC (Criterion Importance Through Intercriteria Correlation) weighting method and particle swarm optimization (PSO) algorithm to optimize support vector regression (SVR) in order to address the issue of no difference in input weight values in landslide displacement prediction, which affects the accuracy of the model prediction. The model first uses Pearson correlation analysis to select three influencing factors with strong correlation with the model output. Then, the CRITIC weighting method is used to obtain the corresponding weights, and the weighted training set is input into a prediction model based on CRITIC weighting method and PSO-SVR to achieve the prediction of landslide surface displacement. The results show that compared with SVR, PSO-SVR, and prediction models based on entropy weighting and PSO-SVR, this model has good generalization ability. The root mean square error and decision coefficient are reduced by 38.24% and increased by 6.64% compared to the unweighted model, respectively, which can effectively improve prediction accuracy. The prediction effect is better than other comparison models.

    参考文献
    相似文献
    引证文献
引用本文

曾子健,肖慧,徐哈宁,等. 基于CRITIC赋权法与PSO-SVR模型的滑坡地表位移预测研究[J]. 科学技术与工程, , ():

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-09-20
  • 最后修改日期:2024-05-18
  • 录用日期:2024-05-21
  • 在线发布日期:
  • 出版日期:
×
亟待确认版面费归属稿件,敬请作者关注