Estimación del valor comercial de los terrenos de la vereda La Sucia en Medellín mediante la evaluación y selección de algoritmos de inteligencia artificial

Abstract

Currently in Colombia, the figure of the Cadastre is being implemented with the so-called Multipurpose Cadastre, where it is defined as a public service; currently the Agustín Codazzi Geographic Institute is the governing body and the so-called cadastral managers are also enabled. The importance of maintaining a formed and updated cadastre gained strength after the peace agreement was signed in 2016 and the diagnoses presented in the CONPES documents 3859 of 2016 and 3958 of 2019, where it was identified that the National territory presents a delay in terms of cadastral information. This panorama leads to challenges for the national territory, with the methods currently proposed by the Agustín Codazzi Geographic Institute in the determination of homogeneous physical and geoeconomic zones, in these the value per square meter of land is defined in areas of similar characteristics, which although they give optimal results, the investment in resources and time is high. This is why there is a need to adopt methodologies that allow optimizing the variables of time and resources. This study aims to analyze 7 artificial intelligence algorithms, measuring their accuracy in determining the commercial value of the land located in the La Sucia area of ​​Medellín - Antioquia, through training the models, analyzing the results and predicting the value.