地理大数据驱动下的中国地理系统多要素时空特征与灾害预测研究  

Research on the Spatio-Temporal Features of Multiple Elements within China’s Geographic System and Disaster Prediction Driven by Geographic Big Data

作　　者：覃浩 高飞 武新浩 

机构地区：[1]西京学院计算机学院,陕西 西安

出　　处：《自然科学》2025年第2期305-319,共15页Open Journal of Nature Science

摘　　要：随着大数据和人工智能的迅猛发展,地理系统问题在地球科学研究中占据着至关重要的地位。它既蕴含着山川湖海的壮美景观以及气候的动态变迁等自然地理现象,又对人口的分布规律、经济活动的开展以及文化的传承发展等人文地理要素产生着深刻影响,然而,存在着诸多的问题。本文利用地理大数据和数学模型,采用随机森林模型和logic回归检验地形对极端天气的影响,随机森林模型进行复杂的非线性关系刻画不同地理因素与极端天气的关联。采用AHP层次分析法进行土地利用变化特征与结构模型推理和构建,通过分析1990~2020年间中国降水量和土地利用/土地覆被的时空演化特征,探讨地形–气候交互作用对极端天气形成的影响,预测2025~2035年间暴雨灾害脆弱地区,并描述中国土地利用变化的特征与结构。With the rapid development of big data and artificial intelligence, the issues of geographical systems occupy a crucial position in earth science research. It encompasses not only natural geographical phenomena such as the magnificent landscapes of mountains, rivers, lakes and seas and the dynamic changes of climate but also has a profound impact on human geographical elements such as the distribution patterns of population, the conduction of economic activities and the inheritance and development of culture. However, there are numerous problems. In this paper, by utilizing geographical big data and mathematical models, the random forest model and logistic regression are employed to examine the impact of terrain on extreme weather. The random forest model depicts the complex nonlinear relationships between different geographical factors and extreme weather. The Analytic Hierarchy Process (AHP) is adopted to conduct reasoning and construction of the model of the characteristics and structure of land use change. Through analyzing the spatio-temporal evolution characteristics of precipitation and land use/lan

关 键 词：地理大数据 LOGISTIC回归 随机森林 层次分析法 

分 类 号：P208[天文地球—地图制图学与地理信息工程]