The root-Gaussian Cox Process for Spatio-temporal Disease Mapping with Aggregated Data
Author
Zeytu Gashaw Asfaw
Co-author
-
Patrick E. Brown
-
Jamie Stafford
-
-
Conference
64th ISI World Statistics Congress
Format: CPS Paper
Keywords: ems algorithm, gaussian random field, matérn correlation and ar(1), spatial statistics
Session: CPS 49 - Statistics and health II and CPS 89 - Spatial statistics and health
Tuesday 18 July 4 p.m. - 5:25 p.m. (Canada/Eastern)
Abstract
This paper focuses on the analysis of spatial data aggregated in space and time when the
boundaries of geographic regions change over time. This can occur when reported cases of a
health outcome are counted in regions over time and these regions change occasionally. We
extend the spatial root-Gaussian Cox Process (RGCP), which uses the square-root link function
rather than the usual log-link function, to the spatio-temporal case. A simulation study shows the
algorithm can identify a spatio-temporal risk surface, and an analysis of malaria incidence in
India is presented.