Urban spatial vulnerability analysis based on urban systems using support vector machine

Abstract

Urban development drives cities to confront new and ongoing challenges, with vulnerability increasingly becoming a key analytical tool and perspective in urban security research. Given that physical space is the primary context for urban events, it requires deeper examination. Current vulnerability studies are less likely to analyze space as a single element and reveal its impact on vulnerability. This paper aims to construct a novel conceptual framework, elaborate on existing vulnerability theories and their manifestations in urban space, and emphasize the importance of spatial function and morphology in vulnerability analysis. From a systems perspective, the relevant spatial elements are reorganized into functional system and morphological system. By integrating these two spatial systems with vulnerability scenarios, the paper explores how each spatial element influences urban spatial vulnerability. In this study, 19 spatial indicators are selected to construct a spatial vulnerability assessment system, the impact of each indicator on spatial vulnerability is analyzed using the Support Vector Machine model. The analysis concludes that the key factors affecting spatial vulnerability are the shape index, patch cohesion index, topographic relief in the morphological system, and the ratio of open space and land use diversity in the functional system. It was found that spatial morphology not only determines the potential vulnerability of physical space to a certain extent, but also has a more significant effect on vulnerability than spatial function before external perturbations occur. On this basis, this study proposes targeted strategies and suggestions to reduce spatial vulnerability, providing reference information and data support for planning and design work.