Data and Replication Package for "Global Riemann-Zeta FPAS+ζ Inflation Forecasting: Layered Annual, Quarterly and Monthly Validation of a Hybrid Global Macroeconomic Model"

This dataset and replication package support the manuscript “Global Riemann-Zeta FPAS+ζ Inflation Forecasting: Layered Annual, Quarterly and Monthly Validation of a Hybrid Global Macroeconomic Model.” The package provides the structured data tables, author-calculated summary results, formula-audit materials, and replication workbook required to reproduce the annual, quarterly, and monthly validation architecture reported in the manuscript. The empirical framework evaluates a hybrid global inflation-forecasting model in which a structural FPAS baseline forecast is augmented by a normalized Riemann-zeta cyclical correction. The Riemann-zeta component is used only as an analogical and computational cyclical signal on the critical line, not as a proof, test, or verification of the Riemann Hypothesis. The admissible empirical claim is limited to whether the normalized zeta-cycle improves and audits global aggregate inflation forecasting when added to a structural FPAS baseline under layered validation. The package covers the manuscript’s layered empirical design: the main annual validation layer for 1970–2024 (N = 55), the annual robustness layer for 1980–2024 (N = 45), the compact crisis-era annual comparison layer for 2005–2024 (N = 20), the quarterly robustness layer for 2000Q1–2024Q4 (N = 100), and the monthly spectral diagnostics layer for 2005M1–2024M12 (N = 240). It includes the reported FPAS, FPAS+ζ, and ARIMA forecast paths, forecast residuals, RMSE/MAE/sMAPE accuracy metrics, alpha-calibration outputs, zeta-cycle and zeta-deviation tables, correction-sign audits, shock-year deformation rankings, composite pressure indicators, zeta-curvature smoothing scores, F_zeta and W_zeta diagnostic functionals, HMM/regime outputs, spectral diagnostics, robustness summaries, and cross-layer validation tables. The replication workbook also includes a methodology formula audit linking the reported calculations to the manuscript’s methodological equations. Supporting files include a data dictionary, table catalog, README, codebook, data availability statement, data status note, Zenodo metadata file, citation file, checksum file, and a reproduction script for selected reported metrics. The dataset should be interpreted as a reported results replication and calculation package. It consolidates the author-calculated tables and validation outputs used in the manuscript and provides traceability from the manuscript tables to structured machine-readable files. External verification against primary raw data providers such as the World Bank, IMF, BIS, FRED, or other official statistical sources requires the separate inclusion of those raw source downloads. Therefore, the package is designed to support transparency, reproducibility of reported calculations, and auditability of the manuscript’s layered validation results, while preserving the study’s stated claim boundaries.