Gold-Silver & Geopolitical Risk

This dataset contains the compiled, structured, and feature-engineered daily time-series data used to model the non-linear transmission channels and risk spillovers between precious metals (gold and silver) and geopolitical risk dimensions. The data supports a triple-target XGBoost machine learning architecture integrated with SHAP (Shapley Additive exPlanations) to evaluate system connectedness across different conditional distributions (Mean, Median Normal, and Extreme Crisis states). The final synchronized dataset ready for machine learning estimation consists of 10,324 rows.