SX-GeoTree: Self-eXplaining Geospatial Regression Tree Incorporating the Spatial Similarity of Feature Attributions
This work addresses the need for interpretable and spatially coherent models in geospatial machine learning, offering a domain-aware explainability template, though it is incremental as it builds on existing methods like decision trees and SHAP.
The paper tackled the problem of decision trees struggling with spatial dependence and unstable explanations by introducing SX-GeoTree, a self-explaining geospatial regression tree that integrates impurity reduction, spatial residual control, and explanation robustness, resulting in competitive predictive accuracy (within 0.01 R² of decision trees) and improved attribution consensus (e.g., modularity: Fujian 0.19 vs 0.09).
Decision trees remain central for tabular prediction but struggle with (i) capturing spatial dependence and (ii) producing locally stable (robust) explanations. We present SX-GeoTree, a self-explaining geospatial regression tree that integrates three coupled objectives during recursive splitting: impurity reduction (MSE), spatial residual control (global Moran's I), and explanation robustness via modularity maximization on a consensus similarity network formed from (a) geographically weighted regression (GWR) coefficient distances (stimulus-response similarity) and (b) SHAP attribution distances (explanatory similarity). We recast local Lipschitz continuity of feature attributions as a network community preservation problem, enabling scalable enforcement of spatially coherent explanations without per-sample neighborhood searches. Experiments on two exemplar tasks (county-level GDP in Fujian, n=83; point-wise housing prices in Seattle, n=21,613) show SX-GeoTree maintains competitive predictive accuracy (within 0.01 $R^{2}$ of decision trees) while improving residual spatial evenness and doubling attribution consensus (modularity: Fujian 0.19 vs 0.09; Seattle 0.10 vs 0.05). Ablation confirms Moran's I and modularity terms are complementary; removing either degrades both spatial residual structure and explanation stability. The framework demonstrates how spatial similarity - extended beyond geometric proximity through GWR-derived local relationships - can be embedded in interpretable models, advancing trustworthy geospatial machine learning and offering a transferable template for domain-aware explainability.