Scholarlink

<p>Liquefaction-induced ground failure poses a significant threat to infrastructure in seismically active regions, such as the Indo-Gangetic Plain, necessitating accurate predictive models for effective risk mitigation. This study proposes a novel <i>Kernel Extreme Learning Machine (KELM)</i> approach to enhance the prediction of liquefaction probability by effectively capturing the complex nonlinear relationships between geotechnical and seismic parameters. Utilizing key input variables, such as standard penetration test values, plasticity index, fines content, peak ground acceleration, and cyclic stress ratio, the model offers improved predictive performance compared to conventional machine learning techniques. A comprehensive evaluation of KELM against established models, such as <i>extreme learning machines</i>, <i>artificial neural networks</i>, <i>support vector machines</i>, <i>random forests</i>, and <i>extreme gradient boosting</i> demonstrates its superior accuracy, robustness, and computational efficiency. The proposed model provides a practical and reliable tool for geotechnical engineers to assess liquefaction susceptibility, thereby enabling informed infrastructure planning and resilience strategies in high-risk zones. This study aligns with the United Nations Sustainable Development Goals (SDGs) by contributing to the development of resilient infrastructure (SDG 9), sustainable urban planning (SDG 11), and climate resilience (SDG 13). The findings underscore the potential of KELM in enhancing disaster preparedness and optimizing resource allocation for sustainable development in liquefaction-prone areas.</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

Kernel-based machine learning model for liquefaction probability prediction and its application to the seismically active indo-gangetic plain

  • 期刊论文

  • 发表日期：

    2025年2月24日

查看全文
  • Sufyan Ghani

摘要

Liquefaction-induced ground failure poses a significant threat to infrastructure in seismically active regions, such as the Indo-Gangetic Plain, necessitating accurate predictive models for effective risk mitigation. This study proposes a novel Kernel Extreme Learning Machine (KELM) approach to enhance the prediction of liquefaction probability by effectively capturing the complex nonlinear relationships between geotechnical and seismic parameters. Utilizing key input variables, such as standard penetration test values, plasticity index, fines content, peak ground acceleration, and cyclic stress ratio, the model offers improved predictive performance compared to conventional machine learning techniques. A comprehensive evaluation of KELM against established models, such as extreme learning machines, artificial neural networks, support vector machines, random forests, and extreme gradient boosting demonstrates its superior accuracy, robustness, and computational efficiency. The proposed model provides a practical and reliable tool for geotechnical engineers to assess liquefaction susceptibility, thereby enabling informed infrastructure planning and resilience strategies in high-risk zones. This study aligns with the United Nations Sustainable Development Goals (SDGs) by contributing to the development of resilient infrastructure (SDG 9), sustainable urban planning (SDG 11), and climate resilience (SDG 13). The findings underscore the potential of KELM in enhancing disaster preparedness and optimizing resource allocation for sustainable development in liquefaction-prone areas.