Artificial neural networks in the application of the growth of the urban sprawl

Abstract

The objective of this work is the use of artificial neural networks and cellular automata to support urban planning decisions in Mexico. We propose an automated model that predicts vertical urban growth, using socio-economic and geographic factors. A multidisciplinary model is presented that manages artificial neural networks, cellular automata, spatial analysis methods, image processing that allow different scenarios of urban growth to be projected and simulated. All of this is built into QGIS through the Python programming language. The model is tested in Mexican cities such as Mexico City, Guadalajara and Monterrey during the years 2015-2020. Reliability ranges from 72% to 76% were obtained, validated by: i) the average number of projected skyscrapers, ii) Position using the Kappa index, and iii) Value in the image using the Jaccard index. With this we propose a technique that allows better informed decisions for urban planning and anticipate new infrastructure needs, projections and regulations.