基于机器学习模型的我国滑坡易发性评价方法综述OA
Review on Landslide Susceptibility Assessment Methods in China Based on Machine Learning Models
[研究目的]滑坡易发性评价是实现滑坡风险管控的前提,对保障山区人民生命财产安全具有重要意义.机器学习模型凭借强大的非线性拟合、高维度数据挖掘和自适应学习等能力,已逐步取代传统评价方法,成为滑坡易发性评价的主要研究方法.[研究方法]本文通过文献梳理,对机器学习方法在我国滑坡易发性评价中的研究进展进行系统总结,梳理开展滑坡易发性评价的技术流程,重点分析数据准备和预处理、机器学习模型类型和组合优化、精度验证与易发性制图等核心环节.[研究结果]研究结果表明,机器学习模型在我国滑坡易发性评价中得到广泛应用,集成模型是目前主流方法之一,取得了良好应用效果.但是在数据层面、模型层面和应用层面,机器学习模型进行滑坡易发性评价时还存在数据和指标因子体系化、模型可释性和普适性、易发性区划图应用程度不足等问题,值得下一步深入研究.本文可为后续机器学习模型在滑坡易发性评价领域的深入研究与实际应用,提供全面的知识框架和参考依据.
[Objective]Landslide susceptibility evaluation is a prerequisite for landslide risk management and control,and it is of great significance for safeguarding the life and property safety of residents in mountainous areas.With powerful capabilities of nonlinear fitting,high-dimensional data mining,and adaptive learning,machine learning models have gradually replaced traditional evaluation methods and become the dominant research method for landslide susceptibility assessment.[Methods]Through literature review,this paper systematically summarizes the research progress of machine learning methods in landslide susceptibility evaluation in China,sorts out the technical workflow of landslide susceptibility evaluation,and focuses on analyzing the core procedures including data preparation and preprocessing,classification and combinatorial optimization of machine learning models,accuracy verification,and susceptibility mapping.[Results]The research results show that machine learning models have been widely applied in domestic landslide susceptibility evaluation,and ensemble models have become one of the mainstream methods with favorable application performance.However,there are still deficiencies in the systematization of data and indicator systems,model interpretability and generalizability,as well as the interdisciplinary application of susceptibility zoning maps when machine learning models are adopted for landslide susceptibility evaluation,which need to be further explored in future studies.This paper can provide a comprehensive theoretical framework and reference basis for the in-depth research and practical application of machine learning models in the field of landslide susceptibility evaluation.
闵承华;张和军;邬学农
湖北煤炭地质一二五队,湖北 宜昌 443003河南省地质局地质灾害防治中心,河南 郑州 450000湖北省地质局第八地质大队,湖北 襄阳 441000
天文与地球科学
机器学习滑坡易发性评价综述
machine learninglandslidesusceptibility evaluationreview
《华南地质》 2026 (4)
561-572,12
兴山县卫生健康局兴山县疾病预防控制中心及实验室能力建设等基坑支护设计项目(A4200002714001058)
评论