Validation of an urban growth prediction model in Quito (Ecuador) built by using weights of evidence and cellular automata

This article analyzes the degree of prediction achieved by an urban growth spatial model based on (1) the historical behavior of formal and informal settlements in the peripheries and (2) the predictive variables of the phenomenon using geospatial techniques. The case study is the north-eastern periphery of the city of Quito (Ecuador), a space that has undergone a great transformation in recent years, is taken as a case study. At first, the spatial model is constructed from satellite images, complemented by analysis of predictive variables of the phenomenon under weights of evidence and cellular automata. Subsequently, the analysis focuses on the validation of the model, through multi-size windows with constant decay function for spatial pattern analysis with DinamicaEgo (R). As a result, the projections generated from the model show co-occurrences consistent with reality and a high percentage of validation.