Scholarlink

<p>This study maps landslide susceptibility in the Gangolihat and Pithoragarh tehsils, Pithoragarh, Uttarakhand, using the Analytical Hierarchy Process (AHP) and Support Vector Machine (SVM). Seven conditioning factors—elevation, slope, aspect, curvature, normalized difference vegetation index (NDVI), land use land cover (LULC), and rainfall—were derived from multi-source datasets and integrated with a verified inventory of 104 landslides. AHP weights, refined through Risk Ratio and Pearson correlation, achieved high consistency (Consistency Ratio, <i>CR</i> = 0.0048). The SVM model, outperformed AHP in predictive accuracy (Area Under Curve, AUC = 0.889 vs. 0.858) and concentrated over 56% of landslides in the highest susceptibility zone. The findings showcase SVM’s superiority for landslide susceptibility mapping, with AHP offering a heuristic alternative for decision support.</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

Landslide Susceptibility Mapping of Pithoragarh, India Using Analytical Hierarchy Process and Support Vector Machine

  • 期刊论文

  • 发表日期：

    2026年1月9日

查看全文
  • Pranshu Vardhan,
  • Rakesh Kumar,
  • Suneet Kaur

摘要

This study maps landslide susceptibility in the Gangolihat and Pithoragarh tehsils, Pithoragarh, Uttarakhand, using the Analytical Hierarchy Process (AHP) and Support Vector Machine (SVM). Seven conditioning factors—elevation, slope, aspect, curvature, normalized difference vegetation index (NDVI), land use land cover (LULC), and rainfall—were derived from multi-source datasets and integrated with a verified inventory of 104 landslides. AHP weights, refined through Risk Ratio and Pearson correlation, achieved high consistency (Consistency Ratio, CR = 0.0048). The SVM model, outperformed AHP in predictive accuracy (Area Under Curve, AUC = 0.889 vs. 0.858) and concentrated over 56% of landslides in the highest susceptibility zone. The findings showcase SVM’s superiority for landslide susceptibility mapping, with AHP offering a heuristic alternative for decision support.