Modeling the spatial relationship between bike-sharing stations and urban centrality using geographical weight variables
This study explores a new approach for including spatial characteristics in machine learning models based on a kernel function in a station-based bike-sharing (SBBS) dataset. On the basis of existing research on geographically weighted statistical methods, we propose a method for transforming spatia...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-08-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843225003978 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This study explores a new approach for including spatial characteristics in machine learning models based on a kernel function in a station-based bike-sharing (SBBS) dataset. On the basis of existing research on geographically weighted statistical methods, we propose a method for transforming spatial characteristics into geographical weight variables (GWVs). Based on the Xuzhou SBBS dataset, machine learning models with GWVs achieve results that are superior to those resulting from most of the spatial regression models and baseline machine learning models with this SBBS dataset. The results show that the GWV approach is suitable for datasets for which regression models have difficulty capturing the spatial features and that this approach can enhance the spatial features for such datasets through the adjustment of the kernel function to obtain better regression results. |
|---|---|
| ISSN: | 1569-8432 |