Hybrid Intelligent System for Leveraging Georeferenced Data and Knowledge

Abstract

This work presents a process for developing an intelligent hybrid system designed to effectively leverage georeferenced data and expert knowledge. The effectiveness of this approach is demonstrated in this work through a specific case study, using the proposed system to achieve a powerful tool for mineral prospectivity. The system consists of three main phases: knowledge and valuable data acquisition, modeling, and results representation using prospectivity heat maps. In the initial step, the recovery and representation of expert knowledge for the case of study was conducted. This system design was tested in the Almadén Mercury Mining District, it involved interviewing expert geologists with ages of experience in the area. Afterwards, the gathering of georeferenced data was carried out to enrich the dataset. Following this phase, the modelling was done, first, using unsupervised techniques to unveil the underlying structure and patterns of the information. Later, employing supervised learning and knowledge representation techniques to enhance the results. In the final step, prospectivity maps were created to represent the achieved results to help in decision making.