Speaker
Description
With the expansion of urban scales and the increasing intricacy of urban population composition, the organization of urban functions and the formation of sub-markets have become increasingly diverse and complex, driven by the growing diversity of urban living demands. Traditional land use strategies are increasingly challenged, and the interplay effects among multiple-function combinations are deemed to play a crucial role. The emergence of Geo AI and explainable AI has provided promising approaches to deciphering the underlying patterns and emergent phenomena in complex urban functional organizations. This study employs an explainable graph neural networks model to uncover the interactions among urban functions along the spatial structure. It conducts a spatially explicit analysis of the importance differences of various functional elements based on the connectivity of the road street network, revealing to urban planners which functional combinations are essential and which can be substituted. Additionally, this study proposes a working framework for integrating the expertise of urban professionals with machine intelligence, playing a key role in the future intelligent development of cities.
| Keywords | Explainable Neural Network; GeoAI; functional organization; |
|---|---|
| Best Congress Paper Award | Yes |
