Aplicación de geoestadística para verificación de litologías mediante la integración de datos geoquímicos y litológicos en la zona Alamala-Catamayo

Abstract

Abstract: This research addresses geological modeling that allows for resource evaluation, the characterization of structures in the mining and oil fields, and the visualization and analysis of the spatial distribution of lithologies present in the study area. This thesis was based on a stochastic model, as it allows for the generation of more in-depth and detailed modeling maps. The applied method of geostatistics is essential for this type of research, allowing for the verification of the spatial distribution of lithologies present in the study area, the Alamala-Catamayo sector, based on the collection of geochemical and lithological data, including chemical composition. Finally, to achieve the aforementioned primary objective, a statistical analysis, multivariate analysis, 17 variograms for each element (variographic analysis), an initial geological map were generated from the collection of geochemical data generated from X-ray fluorescence by previous thesis students, and the predictive map was generated using machine learning, applying the supervised learning method with the "Random Trees with classification and regression" algorithm. This algorithm yielded a confusion matrix indicating the initial values ​​and predictive values ​​after the process was performed.