A functional log-Gaussian Cox approach to predict disease mortality patterns in space and time

A. Torres Signes, M. P. Frías Bustamante, M. D. Ruiz-Medina

A new approach is presented for the statistical analysis of disease mortality patterns in space and time, based on log-Gaussian Cox processes in function spaces. Specifically, this approach involves two steps: (i) The functional parameter estimation of the spatial log-intensity, taking values in a separable Hilbert space; (ii) The functional plug-in prediction of the values of the underlying log-Gaussian Cox process on a time interval, at each spatial region of interest. The results derived are applied to spatial functional prediction of respiratory disease mortality over the Spanish provinces in the Iberian Peninsula.

Keywords: FDA tools Functional Log-Gaussian Cox Processes Mortality Patterns Spatial Functional Statistics.

A. Bücher, H. Dette, F. Heinrichs

D. C. Gamboa Pinilla, A. M. Alonso, D. Peña

J. D. Bermúdez Edo, A. Rubio Fornés, E. Vercher González