Abstract

ROCCO S, Claudio M. Analysis of the residential electric consumption in a venezuelan region using a geographically weighted regression approach. Rev. Fac. Ing. UCV [online]. 2014, vol.29, n.2, pp.7-21. ISSN 0798-4065.

In this paper, Spatial Data Analysis (SDA) and Geographically Weighted Regression (GWR) models are used for quantifying the effects of geo-referenced variables related to the residential electric consumption in a Venezuelan region with several zones. SDA is a statistical technique used to analyze spatial data and able to detect spatial autocorrelation (SA) and spatial heteroscedasticity (SH) effects, which invalidate classical regression model (CRM). GWR allows the efficient quantification of functional relationship in every zone, because it is able to correctly model the spatial effects. The analysis for the selected region suggests the presence of SA and SH. The regression model based on GWR is better that the CRM since it has a higher explanatory capability and allows to quantifying the different effects of the independent variables in each zone.

Keywords : Electrical consumption; Geographically weighted Regression; LISA maps; Moran Index; Spatial data.