Scholarlink

<p>The northern part of the Xiaojiang fault zone was chosen as the research location, and a support vector machine (GEO-SSA-SVM) model optimized by GEO and the Sparrow search algorithm (SSA) was created with the slope unit as the evaluation unit. Seven hundred eighty-four landslide points’ worth of data were gathered via field research and the analysis of remote sensing images. Compounded with the GEO thresholds <i>q</i>(&gt; 0.0179) and <i>p</i>(&lt; 0.1225), eleven important factors were chosen as the landslide susceptibility evaluation criteria. The SSA approach increases the model’s capacity for generalization and prediction accuracy by fine-tuning the parameters of the SVM model. The findings demonstrate that GEO filtering can greatly increase the SVM model’s prediction accuracy. When landslide susceptibility is predicted, the GEO-SSA-SVM model clearly outperforms other conventional models, with an accuracy of 85.80 and an AUC value of 0.915. It is necessary to conduct landslide susceptibility assessment in order to prevent and reduce disasters. This model offers a fresh viewpoint and approach to assessing the susceptibility of landslides.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluation of landslide susceptibility in the northern section of the Xiaojiang fault zone based on factor optimization

  • 期刊论文

  • 发表日期：

    2025年1月20日

查看全文
  • Minghua Hou,
  • Ying Wang,
  • Xianfu Bai,
  • Renmao Yuan

摘要

The northern part of the Xiaojiang fault zone was chosen as the research location, and a support vector machine (GEO-SSA-SVM) model optimized by GEO and the Sparrow search algorithm (SSA) was created with the slope unit as the evaluation unit. Seven hundred eighty-four landslide points’ worth of data were gathered via field research and the analysis of remote sensing images. Compounded with the GEO thresholds q(> 0.0179) and p(< 0.1225), eleven important factors were chosen as the landslide susceptibility evaluation criteria. The SSA approach increases the model’s capacity for generalization and prediction accuracy by fine-tuning the parameters of the SVM model. The findings demonstrate that GEO filtering can greatly increase the SVM model’s prediction accuracy. When landslide susceptibility is predicted, the GEO-SSA-SVM model clearly outperforms other conventional models, with an accuracy of 85.80 and an AUC value of 0.915. It is necessary to conduct landslide susceptibility assessment in order to prevent and reduce disasters. This model offers a fresh viewpoint and approach to assessing the susceptibility of landslides.