A Multivariate Spatial Autocorrelation Statistic, with an Analysis of Noncommunicable Disease-Related Mortality Spatial Clusters.
Conference
64th ISI World Statistics Congress
Format: CPS Abstract
Session: CPS 10 - Disease and mortality modelling
Monday 17 July 8:30 a.m. - 9:40 a.m. (Canada/Eastern)
Abstract
This paper is concerned with the development and application of a new measure of multivariate spatial autocorrelation. The new measure is based on the techniques of canonical correlation analysis. The usefulness of the proposed multivariate statistics is assessed by simulation studies and an application to the multivariate spatial autocorrelation analysis of noncommunicable disease-related mortality rates in South Africa. The estimates of multivariate spatial correlations were unbiased under our proposed methodology. In the application, we identify joint local spatial clusters of the multiple mortality rates.