Abstract
ABSTRACTThe paper designs an automated valuation model to predict the price of residential property in Coventry, United Kingdom, and achieves this by means of geostatistical Kriging, a popularly employed distance-based learning method. Unlike traditional applications of distance-based learning, this papers implements non-Euclidean distance metrics by approximating road distance, travel time and a linear combination of both, which this paper hypothesizes to be more related to house prices than straight-line (Euclidean) distance. Given that – to undertake Kriging – a valid variogram must be produced, this paper exploits the conforming properties of the Minkowski distance function to approximate a road distance and travel time metric. A least squares approach is put forth for variogram parameter selection and an ordinary Kriging predictor is implemented for interpolation. The predictor is then validated with 10-fold cross-validation and a spatially aware checkerboard hold out method against the almost exclusively employed, Euclidean metric. Given a comparison of results for each distance metric, this paper witnesses a goodness of fit () result of 0.6901 ± 0.18 SD for real estate price prediction compared to the traditional (Euclidean) approach obtaining a suboptimal value of 0.66 ± 0.21 SD.
Highlights
By 2030, investable real estate is expected to have grown by more than 55%, amounting to a UK residential market value of £9.145 trillion (IPF 2017)
Nonmetric pairwise road distance and travel time matrices are highly unlikely to produce a valid variogram and covariance function. This puts forth a set of valid Minkowski distance metrics which are proven to better approximate restricted road distance and travel time across the Coventry (United Kingdom) road network compared with a Euclidean distance
This paper hypothesizes that house prices are related to a more complex structural network relating to road distance and travel time; we introduce an approximate road distance and travel time metric using the Minkowski distance function for a valid house price Kriging predictor (Matheron 1963; Cressie 1990)
Summary
By 2030, investable real estate is expected to have grown by more than 55%, amounting to a UK residential market value of £9.145 trillion (IPF 2017). Machine-learning algorithms, under the name of automated valuation models (AVMs), exploit these data to reliably understand the value of real estate over large areas where market behavior may differ significantly. One must ensure a valid positive definite covariance and conditionally negative definite variogram which cannot be guaranteed with any non-Euclidean distance functions (Curriero 2005, 2006) For this reason, nonmetric pairwise road distance and travel time matrices are highly unlikely to produce a valid variogram and covariance function. Nonmetric pairwise road distance and travel time matrices are highly unlikely to produce a valid variogram and covariance function This puts forth a set of valid Minkowski distance metrics which are proven to better approximate restricted road distance and travel time across the Coventry (United Kingdom) road network compared with a Euclidean distance.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
