Validación y análisis del uso del método Gradient Boosting en la estimación del valor de terreno para predios con áreas superiores a una (1) hectárea, en zona rural del municipio La Mesa, Cundinamarca

Abstract

The study entitled "Validation and analysis of the use of the Gradient Boosting method in the estimation of land value for properties with areas greater than one (1) hectare, in rural area of La Mesa municipality, Cundinamarca" focuses on evaluating the effectiveness of the Gradient Boosting method in the estimation of land value in rural areas for properties with agricultural characteristics. The Gradient Boosting method is a machine learning technique that has proven to be effective in solving complex problems in various fields, including the estimation of real estate value. The work was developed from information obtained from the La Mesa municipality, Cundinamarca and focused on properties with areas greater than one hectare. Based on the available information, some variables were defined that formed the model, including: water availability, road influence, economic destination, class of homogeneous land area, land area, slope, climate unit, potential value and agricultural class. The Gradient Boosting method was applied using the R studio software to this data to estimate the land value of these properties and was compared to the actual values. The results of the study showed that the Gradient Boosting method is an effective tool for the estimation of land value in rural areas, with high accuracy in estimation. Additionally, the variables that have a significant impact on the value of agricultural properties were identified.