Spatially-Weighted Factor Analysis for Extraction of Source-Oriented Mineralization Feature in 3D Coordinates of Surface Geochemical Signal

Prediction of gold mineralization zones using spatial techniques and geophysical data: A case study of the Josephine prospecting licence, NW Ghana

“…In the case of the data-driven methods, the weight of each of the geoscientific criteria to be used in the predictive modeling is determined by assessing how they spatially correlate with respect to known locations of the mineral occurrences within the study area [ 46 ]. The use of data-driven methods in predicting the spatial occurrence of a natural resource or a geohazard is mostly carried out by the weight of evidence [ 47 , 74 ], frequency ratio [ 16 , 53 , 69 ], weighting factor [ 9 , 23 , 35 , 36 ], statistical information [ 35 , 52 ], information value [ 6 , 27 ], shannon entropy [ 8 , 77 ], certainty factor [ 56 , [80] , [80a] ], evidence belief function [ 40 , 62 ], neural networks [ 55 , 57 ], logistic regression [ 29 , 58 ], support vector machine [ 28 , 82 ], and random forest [ 65 , 80 ] techniques. It should also be emphasized that, data-driven methods do not work well in situations where the known locations of the sought-after mineral is limited or absent.…”