An Improved Data-Driven Multiple Criteria Decision-Making Procedure for Spatial Modeling of Mineral Prospectivity: Adaption of Prediction–Area Plot and Logistic Functions

Assigning realistic weights to targeting criteria in order to synthesize various geo-spatial datasets is one of the most important challenging tasks for mineral prospectivity modeling (MPM). Techniques for multiple criteria decision-making (MCDM), like MPM, are deeply concerned with combining a large-scale exploration dataset into a single evaluation model for localizing prospects of a certain deposit type. In this paper, we develop the data-driven TOPSIS procedure, as a GIS-based MCDM technique for MPM. Because weighting and integrating various exploration evidence layers are influenced by intricacy and vagueness of ore mineralization process, imprecise selection of targeting criteria may reduce the possibility of exploration success. To address this problem, we applied prediction–area plot for prioritizing, recognizing and weighting efficient and inefficient targeting criteria. In addition, normalized density (Nd) index was then used for assigning significant weights to fractal-based discretized classes of each targeting criterion. After recognition of efficient and inefficient targeting criteria, data-driven TOPSIS procedure was adapted based on participation of only efficient targeting criteria as well as all targeting criteria for porphyry-Cu prospectivity in Varzaghan district, NW Iran. For quantitative assessment, a success rate curve for each of the two prospectivity models generated in this study was drawn. The results prove the superiority of the predictive model based on using efficient targeting criteria.