Exploring spatial patterns of mortality in Europe using functional spatial principal components for areal data - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Preprints, Working Papers, ... Year : 2021

Exploring spatial patterns of mortality in Europe using functional spatial principal components for areal data

Abstract

We examined the spatial autocorrelation of mortality rates for 28 European countries, with data from the Human Mortality Database (HMD) using spatial associations in the context of functional areal data. We developed a functional Moran's I statistic which is the first of its kind in the functional data analysis framework to determine spatial autocorrelation and spatial PCA for areal data. These data were converted to functions before performing the classical and spatial PCA. Results showed the existence of spatial autocorrelation between neighbouring countries (using K-nearest neighbours (KNN) and contiguity neighbours) based on the functional Moran's I statistic applied on the functional PCA approximation. However, no strong correlation was displayed when the scores of the classical PCA which ignored spatial information, was considered. This work proved the existence of spatial dependency in mortality rates of neighbouring countries in Europe and showed that the functional Moran's I statistic is a powerful tool in measuring spatial dependency.
No file

Dates and versions

hal-03133827 , version 1 (07-02-2021)

Identifiers

  • HAL Id : hal-03133827 , version 1

Cite

Dharini Pathmanathan, Alaa Ali Hassan, Alaa Hassan, Sophie Dabo-Niang. Exploring spatial patterns of mortality in Europe using functional spatial principal components for areal data. 2021. ⟨hal-03133827⟩
104 View
0 Download

Share

Gmail Facebook X LinkedIn More