SAS macro programs for geographically weighted generalized linear modeling with spatial point data: applications to...

Area of Science:

• Spatial statistics
• Geographic Information Systems (GIS)
• Statistical software development

Background:

• Growing interest in spatial non-stationarity necessitates advanced analytical tools.
• Existing specialized software often lacks integration with established statistical environments like SAS.
• Geographically Weighted Generalized Linear Modeling (GWGLM) is a powerful technique for spatial analysis.

Purpose of the Study:

• To develop SAS macro programs for integrated GWGLM analysis.
• To expand the capabilities of GWGLM within the SAS environment.
• To provide a flexible and powerful tool for exploring spatial non-stationarity.

Main Methods:

• Development of a novel set of SAS macro programs.
• Integration of GWGLM principles with SAS functionalities.
• Implementation of enhanced kernel weighting functions and bandwidth selection methods.
• Empirical validation using three distinct case studies.

Main Results:

• The developed SAS macro programs offer expanded kernel weighting functions.
• Users gain improved control over bandwidth selection processes.
• The GWGLM framework is successfully embedded within the SAS environment.
• Demonstrated utility and advantages through empirical examples.

Conclusions:

• The SAS macro programs effectively bridge the gap between GWGLM and SAS.
• The developed tools facilitate more sophisticated spatial non-stationarity modeling.
• This integration offers significant potential for future research in complex spatially varying coefficient models across disciplines.