Applied machine learning predictive modelling in regional spatial data analysis problem

Abstract

Urban and Regional Studies deal with large tables of spatial data obtained from censuses and surveys. It is necessary to simplify the huge amount of detailed information in order to extract the main trends. The main aim of this article is to present and compare machine learning models in spatial data analysis problem. As an example of spatial data modelling we draw upon public-domain data about California housing values. We use a variety of regression modelling techniques, showing how additional information about location (longitude and latitude) can contribute to the analysis. The data comprise observations of housing values, economic covariates, and longitude and latitude. We follow Pace and Barry (1997) in defining response and explanatory variables for a linear regression model.