Exploring Nonlinear and Interactive Influences of Built Environment on Metro‐To‐Bus Transfer Behavior

As buses increasingly emerge as a pivotal feeder mode for metro, unraveling the transfer behavior between metro and bus is crucial for enhancing the intermodal integration and improving the attractiveness of the whole transit system. Employing Shanghai central city as a case study, this study adopted Random Forest to unravel the complex (nonlinear and interactive) effects between metro‐to‐bus transfer ridership and built environment variables across four distinct weekday time periods (morning peak, afternoon off‐peak, afternoon peak and late‐night hours). Our empirical analysis highlights the differential influences of key built environment factors as well as their nonlinear and threshold patterns on transfer ridership across these varying temporal contexts. During peak hours, transfer time (26.16% of RI value) and bus route (21.49% of RI value) significantly influence passengers' transfer behavior, while the impact of land‐use mix (13.45% of RI value) becomes more prominent during off‐peak hours. Furthermore, this research highlights the complex and temporal‐dynamic interactions between built environment factors on transfer ridership: the intensity of interactions between built environment variables is markedly higher during peak hours compared to off‐peak hours. This study provides policymakers with an in‐depth understanding of the temporal heterogeneity of transfer behavior, thereby helping to develop an efficient transit system.