Search

Techniques for GIS-based multicriteria decision-making (GIS-MCDM), like mineral prospectivity mapping (MPM), are concerned with combining information from several criteria into a single evaluation model to solve certain problems in the fields of geosciences. In this paper, we introduce the geometric average method for MPM, as a GIS-MCDM approach, and demonstrate its advantage over the expected value MPM method. The comparative analysis shows that the geometric average MPM method yields better prediction of mineral prospectivity and it overcomes the limitation of the expected value method in terms of using spatial evidence values with the same unit. [ABSTRACT FROM AUTHOR]

There are methods of mineral prospectivity mapping whereby, besides assignment of weights to classes of evidence in an evidential map, every evidential map is also given a weight based on expert opinion. In this regard, evaluating the relative importance of every evidential map derived from particular spatial data sets is a highly subjective exercise and the assignment of meaningful weights to evidential maps usually involves a trial-and-error procedure. In this paper, we used a prediction–area (P–A) plot and normalized density to estimate weights of every evidential map. The method of P–A plot is a data-driven way, rather than using expert opinion, to evaluate and weight evidential maps. [ABSTRACT FROM AUTHOR]

Complexities of geological processes portrayed as certain feature in a map (e.g., faults) are natural sources of uncertainties in decision-making for exploration of mineral deposits. Besides natural sources of uncertainties, knowledge-driven (e.g., fuzzy logic) mineral prospectivity mapping (MPM) is also plagued and incurs further uncertainty in subjective judgment of analyst when there is no reliable proven value of evidential scores corresponding to relative importance of geological features that can directly be measured. In this regard, analysts apply expert opinion to assess relative importance of spatial evidences as meaningful decision support. This paper aims for fuzzification of continuous spatial data used as proxy evidence to facilitate and to support fuzzy MPM to generate exploration target areas for further examination of undiscovered deposits. In addition, this paper proposes to adapt the concept of expected value to further improve fuzzy logic MPM because the analysis of uncertain variables can be presented in terms of their expected value. The proposed modified expected value approach to MPM is not only a multi-criteria approach but it also treats uncertainty of geological processes a depicted by maps or spatial data in term of biased weighting more realistically in comparison with classified evidential maps because fuzzy membership scores are defined continuously whereby, for example, there is no need to categorize distances from evidential features to proximity classes using arbitrary intervals. The proposed continuous weighting approach and then integrating the weighted evidence layers by using modified expected value function, described in this paper can be used efficiently in either greenfields or brownfields. [ABSTRACT FROM AUTHOR]