摘要
洪涝灾害是最为严重的自然灾害之一,相关研究众多。该文采用隐含狄利克雷分布主题模型和词频-逆文档频率算法,对中文洪涝灾害文献进行深入的文本分析,旨在揭示洪涝灾害研究整体特征和时空异质性。结果表明:①经过主题挖掘,可将现有研究归为洪涝灾害风险与响应、水文模型和气候、城市洪涝管理、生态保护和社会发展以及农业影响等五大维度。②基于词频-逆文档频率算法对研究时空异质性进行辨析;从时间维度来看,研究可分为防控意识觉醒期、系统研究发展期和多元研究迸发期等三个阶段;从空间视角来看,自然社会背景对研究重点等具有基础性影响。基于以上分析,对洪涝灾害研究未来发展方向进行研判。
Flood disasters are among the most severe natural disasters,attracting extensive research.By employing the Latent Dirichlet Allocation topic model and the Term Frequency-Inverse Document Frequency algorithm,an in-depth textual analysis of Chinese flood disaster literature is done,aiming to reveal the overall characteristics and spatio-temporal heterogeneity of flood disaster research.The results show that:(1)Through topic mining,existing studies can be categorized into five main dimensions:flood disaster risk and response,hydrological models and climate,urban flood management,ecological protection and social development,and agricultural impacts.(2)Based on the TF-IDF algorithm,the spatio-temporal heterogeneity of the research is discerned;from a temporal dimension,research can be divided into stages of awareness of prevention and control,period of systematic research development,and era of diversified research eruption;from a spatial perspective,natural and social backgrounds have a fundamental impact on research focuses.Based on this analysis,the future research directions of flood disaster studies is prognosticated.
作者
宋苑震
田健
艾合麦提·那麦提
贺蔚杰
曾坚
SONG Yuanzhen;TIAN Jian;AIHEMAITI Namaiti;HE Weijie;ZENG Jian(School of Architecture,Tianjin University,Tianjin 300000,China)
出处
《灾害学》
CSCD
北大核心
2024年第4期47-54,共8页
Journal of Catastrophology
基金
来源“基于智慧与韧性适灾机理的沿海河口城市雨潮组合风险防控规划理论”(52378065)
关键词
洪涝灾害
时空异质性
文本分析
主题挖掘
flood disasters
spatio-temporal heterogeneity
textual analysis
topic mining
