Robust geographical detector

• A Robust Geographical Detector (RGD) is proposed to quantify spatial associations. RGD includes a rank function and change point detection-based optimization. eliminates the sensitivity of discretization in stratified heterogeneity. estimates robust power determinants (PD) variables using B-values. provides higher more stable PD than previous GD models. detector (GD) method measure associations determinant value that compares variance data within zones whole study area. Recent studies have implemented diverse fields, such as environmental socio-economic issues. Spatial an essential stage for determining explanatory variables. However, process has been sensitive results. To address this issue, article proposes overcome limitations estimate values B-value. The determines with numerical interval breaks optimization algorithm variance-based detection. In study, nationwide case exploring potential factors nitrogen dioxide (NO 2 ) density industrial regions across Australia, where on both NO are sourced from satellite images remote sensing products Google Earth Engine. Results show can effectively explore maximum between response due algorithm-based zones. addition, RGD-based generally higher, robust, GD-based since guarantee increment increase numbers, which challenge Finally, could provide reliable interpretation finds optimal intervals-based determined by factors. This demonstrates developed model solutions identify geographical