Geographically weighted regression with a non-Euclidean distance\ud metric: a case study using hedonic house price data

摘要

Geographically weighted regression (GWR) is an important local technique for exploring\udspatial heterogeneity in data relationships. In fitting with Tobler’s first law of\udgeography, each local regression of GWR is estimated with data whose influence\uddecays with distance, distances that are commonly defined as straight line or\udEuclidean. However, the complexity of our real world ensures that the scope of\udpossible distance metrics is far larger than the traditional Euclidean choice. Thus in\udthis article, the GWR model is investigated by applying it with alternative, non-\udEuclidean distance (non-ED) metrics. Here we use as a case study, a London house\udprice data set coupled with hedonic independent variables, where GWR models are\udcalibrated with Euclidean distance (ED), road network distance and travel time metrics.\udThe results indicate that GWR calibrated with a non-Euclidean metric can not only\udimprove model fit, but also provide additional and useful insights into the nature of\udvarying relationships within the house price data set.