Explore the Impact Mechanism of Urban Built Environment on Thermal Environment Based on Deep Machine Learning

In recent years, there have been numerous studies on the impact of urban morphological characteristics on the urban thermal environment. However, the majority of these studies have focused on the entire city rather than the specific built-up areas. This study employs a Feed-forward Neural Network (FNN) regression model to maximize the fitting of the relationship between built environment factors and the thermal environment. Additionally, we introduce the Shapley Additive exPlanations model variable explanation method to interpret the impact of built environment morphological elements on the nonlinear relationship with surface temperature. The nonlinear features in the FNN model are revealed by quantitatively analyzing the importance and direction of influence of various built environment variables on the thermal environment.