EXPLORING THE ASSOCIATION BETWEEN BUILT ENVIRONMENT AND URBAN VITALITY USING DEEP LEARNING METHODS

Abstract

Urban vitality is a critical element in the development of cities. The built environment of a city plays a pivotal role in shaping urban vitality. Using Yinchuan City as a case study, this research employs the Baidu heat map to assess urban vitality. Simultaneously, the built environment variables serve as independent variables. We utilize Ordinary Least Squares (OLS), Moran's I, and Geographically Weighted Regression (GWR) models to explore the relationship between urban vitality and the built environment on weekdays and weekends in Yinchuan City. Finally, we apply the Gradient Boosting Decision Tree (GBDT) model to analyze the importance of variables influencing different time periods of urban vitality. The research findings indicate that: (1) The built environment in Yinchuan City significantly influences urban vitality on both weekdays and weekends. (2) There is positive spatial autocorrelation between the built environment and urban vitality on both weekdays and weekends. (3) GWR model analysis reveals that urban vitality on weekdays and weekends exhibits different spatial distribution characteristics. (4) GBDT model analysis indicates that variables influencing urban vitality during weekdays and weekends have different importance rankings. Finally, tailored strategies to enhance urban vitality are proposed for different urban areas and time periods. This study provides crucial reference points for urban planning and sustainable development in Yinchuan City.

Keywords

• Built Environment
• Baidu Heat Map
• Urban Vitality
• Deep Learning
• GBDT