Revisiting Gaussian Process Reconstruction for Cosmological Inference: The Generalized Gaussian Process Framework

Abstract We investigate uncertainties in the estimation of the Hubble constant ( H 0 ) arising from Gaussian process (GP) reconstruction, demonstrating that the choice of kernel introduces systematic variations comparable to those arising from different cosmological models. To address this limitation, we introduce the generalized Gaussian process (Gen GP) framework, in which the Matérn smoothness parameter ν is treated as a free parameter, allowing for data-driven kernel optimization. Using the cosmic chronometer Hubble data, we find that while standard GP with a Λ-cold-dark-matter (ΛCDM) mean function exhibits noticeable reconstruction differences between optimized and marginalized approaches, particularly at z > 1, Gen GP maintains methodological consistency. In Gen GP, slight increases in χ 2 per degree of freedom relative to standard GP, for both the zero-mean and ΛCDM prior mean cases, reflect added flexibility rather than performance degradation. Our results emphasize that a robust cosmological inference requires treating kernel parameters as free variables and implementing full Bayesian marginalization to avoid artificial precision from fixed hyperparameters. As machine learning becomes central to cosmological discovery, the Gen GP framework provides a principled approach to a model-independent inference that properly accounts for methodological uncertainties while maintaining the necessary flexibility for a reliable parameter estimation.