机器学习在国内城市洪涝灾害研究中的进展与热点分析

Progress and Hotspot Analysis of Machine Learning in Domestic Urban Flood Disaster Research

摘要:

机器学习作为一种新兴的技术方法，近年来在城市洪涝灾害研究中的应用价值不断凸显。利用文献计量可视化工具CiteSpace对1986—2024年来国内基于机器学习的城市洪涝灾害研究进行梳理与分析，揭示研究领域的整体发展脉络、研究热点及未来趋势。主要结论：（1）机器学习在国内城市洪涝灾害中的研究成果数量经历平稳—升温—波动—快速上升4个阶段；（2）研究作者和研究机构呈现出一定程度的聚集；发文期刊仅约有一半的比例属于核心期刊，且CSCD与CSSCI期刊占比不高；（3）机器学习在城市洪涝灾害研究中呈现内容多样化特征，在以往洪水预报研究为主逐步向洪涝灾害风险评估为趋势转变。

Abstract:

As a new technical method, machine learning has been increasingly applied in the study of urban flood disaster in recent years. The bibliometric visualization tool CiteSpace is used to sort out and analyze the research on urban flood disaster based on machine learning in China from 1986 to 2024, which reveals the overall development trend, research hotspots and future trends in the research field. The main conclusions are that (1) The number of research achievements of machine learning in domestic urban flood disasters has experienced four stages: steady, warming, fluctuating and rapidly rising. (2) Research authors and research institutions present a certain degree of clustering. Only about half of the published journals belong to the core journals, and the proportion of CSCD and CSSCI journals is not high. (3) The content of machine learning in urban flood disaster research is diversified. In the past, the flood forecasting research has gradually shifted to flood disaster risk assessment.