Geographically weighted regression is a powerful and computationally intensive method to model varying spatial relationships, but it may introduce high local multicollinearity which, if not dealt with properly, leads to misleading inference and unreliable results. We introduced a novel solution to deal with multicollinearity by adopting a more localized and refined approach. This solution is demonstrated by modeling the varying local association of childhood obesity and its risk factors at neighborhood level to target only vulnerable neighborhoods for the prevention and control of obesity locally
Addressing Multicollinearity in Local Modeling of Spatially Varying Relationship using GWR / Rizwan Shahid; Stefania Bertazzon. - ELETTRONICO. - (2017), pp. 0-0. (Intervento presentato al convegno Spatial Knowledge and Information Canada tenutosi a Banff, Canada nel 2017).
Addressing Multicollinearity in Local Modeling of Spatially Varying Relationship using GWR
Stefania Bertazzon
2017
Abstract
Geographically weighted regression is a powerful and computationally intensive method to model varying spatial relationships, but it may introduce high local multicollinearity which, if not dealt with properly, leads to misleading inference and unreliable results. We introduced a novel solution to deal with multicollinearity by adopting a more localized and refined approach. This solution is demonstrated by modeling the varying local association of childhood obesity and its risk factors at neighborhood level to target only vulnerable neighborhoods for the prevention and control of obesity locallyI documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.