Premium
Using Contextualized G eographically W eighted R egression to Model the Spatial Heterogeneity of Land Prices in B eijing, C hina
Author(s) -
Harris Rich,
Dong Guanpeng,
Zhang Wenzhong
Publication year - 2013
Publication title -
transactions in gis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.721
H-Index - 63
eISSN - 1467-9671
pISSN - 1361-1682
DOI - 10.1111/tgis.12020
Subject(s) - weighting , set (abstract data type) , covariate , space (punctuation) , geography , econometrics , computer science , mathematics , programming language , operating system , medicine , radiology
G eographically W eighted R egression ( GWR ) is a method of spatial statistical analysis allowing the modeled relationship between a response variable and a set of covariates to vary geographically across a study region. Its use of geographical weighting arises from the expectation that observations close together by distance are likely to share similar characteristics. In practice, however, two points can be geographically close but socially distant because the contexts (or neighborhoods) within which they are situated are not alike. Drawing on a previous study of geographically and temporally weighted regression, in this article we develop what we describe as contextualized G eographically W eighted R egression ( CGWR ), applying it to the field of hedonic house price modeling to examine spatial heterogeneity in the land parcel prices of Beijing, C hina. Contextual variables are incorporated into the analysis by adjusting the geographical weights matrix to measure proximity not only by distance but also with respect to an attribute space defined by measures of each observation's neighborhood. Comparing CGWR with GWR suggests that adding the contextual information improves the model fit.