“…Thus, the optimal values of the hyperparameters are derived from the maximisation of the probability of the GP to generate our considered set of data, that is implemented [88,[91][92][93] via the minimisation of the GP marginal likelihood, which is similar to the hierarchical Bayesian approach. GP have now been exhaustively used for the reconstruction of cosmological functions, particularly related to the late-time cosmic accelerated expansion observables [75,76,88,89,[94][95][96][97][98][99][100][101][102][103][104][105]. We should remark that although GP are independent from any cosmological model, GP rely on the choice of the kernel function which governs the correlations between distinct points in the GP reconstructed function, and hence its profile.…”