Title: Geographically weighted regression analysis with two-factorial compositional covariates
Authors: Kamila Facevicova - Palacky University Olomouc (Czech Republic) [presenting]
Petra Kynclova - United Nations Industrial Development Organization (Austria)
Karel Macku - Palacky University Olomouc (Czech Republic)
Abstract: An approach is introduced to modeling the relationship between a real variable and a relative structure under the presence of spatial dependence. Spatial statistics provides a wide range of methods for the analysis of data with local variations but these methods are rarely accommodated to work with data of a relative nature. The presented methodology is motivated by the problem of modelling local variations of the relationship between at-risk-of-poverty rates and the structure of the German population aged 30-34, given by gender and the highest attained educational level. Consequently, the geographically weighted regression model with covariates forming a compositional table is introduced and interpretation of regression coefficients is discussed.