Remote sensing-based measurement of Living Environment Deprivation: Improving classical approaches with machine learning

• Published in: PloS one Vol. 12; no. 5; p. e0176684
• Main Authors: Arribas-Bel, Daniel, Patino, Jorge E., Duque, Juan C.
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 02.05.2017; Public Library of Science (PLoS)
• Subjects: Analysis; Artificial intelligence; Cities; Computer and Information Sciences; Data Interpretation, Statistical; Demographics; Earth Sciences; Ecology and Environmental Sciences; Engineering and Technology; England; Historic preservation; Housing conditions; Humans; Land use; Learning algorithms; Least-Squares Analysis; Linear Models; Machine Learning; Medicine and Health Sciences; Physical Sciences; Quality of life; Remote sensing; Remote Sensing Technology - methods; Research and Analysis Methods; Social Sciences; Socioeconomic Factors; Socioeconomics; Spatial analysis; Trends; Urban areas; Urban planning; Urban Population; Variables; Indiana; United Kingdom; Detroit Michigan; United States > US
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0176684

This paper provides evidence on the usefulness of very high spatial resolution (VHR) imagery in gathering socioeconomic information in urban settlements. We use land cover, spectral, structure and texture features extracted from a Google Earth image of Liverpool (UK) to evaluate their potential to predict Living Environment Deprivation at a small statistical area level. We also contribute to the methodological literature on the estimation of socioeconomic indices with remote-sensing data by introducing elements from modern machine learning. In addition to classical approaches such as Ordinary Least Squares (OLS) regression and a spatial lag model, we explore the potential of the Gradient Boost Regressor and Random Forests to improve predictive performance and accuracy. In addition to novel predicting methods, we also introduce tools for model interpretation and evaluation such as feature importance and partial dependence plots, or cross-validation. Our results show that Random Forest proved to be the best model with an R2 of around 0.54, followed by Gradient Boost Regressor with 0.5. Both the spatial lag model and the OLS fall behind with significantly lower performances of 0.43 and 0.3, respectively.