EMS Response Time Analyses for a Rural County Using Geographically Weighted Regression with Different Kernel Weighting Functions
UAH PRC Research Database
International Journal of Statistics and Applications
A Geographically Weighted Regression (GWR) is considered to compare results provided using two different kernel weighting functions: adaptive bi-square kernel and adaptive Gaussian kernel. To provide a baseline reference comparison, resulting data are also considered relative to Global Regression Analysis (GRA) calculations, which are obtained without the inclusion of geographical variability location data. For the analysis, data associated with a total of 214 crash cases for the dates between January 2016 and December 2019 are studied for a rural county in Alabama. Associated crash records are extracted from the Critical Analysis Reporting Environment (CARE) database. Six independent variables, including travel time, time of the day, day of the week, weather, lighting conditions, and crash severity are modeled in regard to their influences on EMS Response Time (ERT). Results from GWR analyses, using both weighting functions, show important quantitative and qualitative differences in regard to coefficient values as each independent variable is individually addressed, especially as the number of considered variables is altered, relative to the addressed variable. Mean square (MS) values associated with GWR Residuals are 276.6 for the adaptive bi-square kernel function and 332.4 for the adaptive Gaussian kernel function. Such differences within ANOVA table data indicate that GWR analysis, with an adaptive bi-square kernel weighting function, often yields improved model performance, relative to GWR with an adaptive Gaussian kernel weighting function. ANOVA table data also evidence improved model performance with the inclusion of geographical variability location data.
Vanga, Sneha R.; Ligrani, Phillip M.; Doustmohammadi, Mehrnaz; and Anderson, Michael, "EMS Response Time Analyses for a Rural County Using Geographically Weighted Regression with Different Kernel Weighting Functions" (2022). PRC-Affiliated Research. 16.