Scholarlink

<p>Floods frequently lead to significant environmental disruption and economic losses. Traditional prediction methods often struggle to address the multifaceted nature of flood events due to data limitations. This study developed a predictive model to assess flood susceptibility for the Shebelle Basin, Somalia, by comparing MCDA with the GBC. A balanced dataset comprising 2,077 flooded and 2,077 non-flooded points was analyzed, with 70% of the dataset used for training and 30% for testing. This supported a detailed assessment of the model’s precision in predicting flood susceptibility. The GBC showed a significant improvement over AHP, achieving accuracy, precision, and an area under the curve (AUC) of 89.41%, 85.65%, and 0.96, respectively, while the AHP achieved an AUC of 0.79. The results highlight the enhanced capability of the GBC to effectively integrate and analyze diverse hydrological data, establishing it as a more reliable tool for flood risk assessment in large regions like the Shebelle River Basin in Somalia. The use of advanced machine learning approaches, such as the GBC algorithm, emerges as a powerful and economically feasible method for comprehensive geospatial modelling of flood susceptibility, offering valuable insights to support emergency response, preparedness, and flood mitigation plans.</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

Assessment of flood susceptibility mapping using integrated multi-criteria decision analysis and machine learning

  • 期刊论文

  • 发表日期：

    2025年12月10日

查看全文
  • A. M. Mohamoud,
  • A. M. Al-Areeq,
  • S. Chowdhury,
  • M. Al-Suwaiyan

摘要

Floods frequently lead to significant environmental disruption and economic losses. Traditional prediction methods often struggle to address the multifaceted nature of flood events due to data limitations. This study developed a predictive model to assess flood susceptibility for the Shebelle Basin, Somalia, by comparing MCDA with the GBC. A balanced dataset comprising 2,077 flooded and 2,077 non-flooded points was analyzed, with 70% of the dataset used for training and 30% for testing. This supported a detailed assessment of the model’s precision in predicting flood susceptibility. The GBC showed a significant improvement over AHP, achieving accuracy, precision, and an area under the curve (AUC) of 89.41%, 85.65%, and 0.96, respectively, while the AHP achieved an AUC of 0.79. The results highlight the enhanced capability of the GBC to effectively integrate and analyze diverse hydrological data, establishing it as a more reliable tool for flood risk assessment in large regions like the Shebelle River Basin in Somalia. The use of advanced machine learning approaches, such as the GBC algorithm, emerges as a powerful and economically feasible method for comprehensive geospatial modelling of flood susceptibility, offering valuable insights to support emergency response, preparedness, and flood mitigation plans.