Bayesian surface regression versus spatial spectral nonparametric curve regression

COVID-19 incidence is analyzed at the provinces of some Spanish Communities during period February-October, 2020. Two infinite-dimensional regression approaches are tested. The first one implemented in framework introduced Ruiz-Medina, Miranda and Espejo (2019). Specifically, a bayesian adopted estimation pure point spectrum temporal autocorrelation operator, characterizing second-order structure surface sequence. second approach formulated context spatial curve regression. A nonparametric estimator spectral density based on periodogram computed to approximate correlation between curves. Dimension reduction achieved by projection onto empirical eigenvectors long-run covariance operator. Cross-validation procedures test performance two functional approaches.