Document Type : Original Research Paper

Authors

• Hamidreza Ramazi ¹
• Moslem Jahantigh ²

¹ Amirkabir University of Technology

² Mine engineering, Amirkabir university of technology, Tehran, Iran

Abstract

The present study aims to compare data-driven and knowledge-driven methods in weighting exploration witness layers in porphyry copper potential in the study area of Shahr-e-Babak, Iran. Initially, eight exploration layers including geochemical, geological, airborne magnetic, and remote sensing data were preprocessed and prepared. Then, these layers were transferred to an equal distance of zero and one using a fuzzy function. The weighting of exploration witness layers in the study area was done using three methods: AHP, random forest algorithm, and area prediction method. Then, the weighted layers were combined using the MABAC multivariate decision-making method and three mineral potential models were formed for the study area. The produced models were compared using the main characteristic curve diagram. It was found that weighting using a data-driven method has more desirable results than a knowledge-driven method. Also, weighting exploration layers by random forest algorithm has more desirable results than the area prediction method. The better results of the random forest algorithm are because in this algorithm, weighting is based on mineralized and non-mineralized points, while in the area prediction method, weighting is only based on outcrops in the study area. This method can be implemented with appropriate accuracy in areas with limited training data, along with knowledge-based methods for weighting exploratory evidential layers.

Keywords

• MABAC
• Random forest
• AHP
• Porphyry
• Shahr-e-Babak

Main Subjects

• Exploitation