R-GeoXNet: A Direction-Aware Relational Graph Neural Network for Regional Mineral Prospectivity Mapping
This study introduces R-GeoXNet, a direction-aware relational graph neural network that integrates geochemical and structural data to effectively model anisotropic geological controls, achieving superior performance in regional polymetallic prospectivity mapping within the Lhasa-Woka area of Tibet compared to existing graph-based baselines.