3D Geophysical Predictive Modeling by Spectral Feature Subset Selection in Mineral Exploration

Several technical challenges are related to data collection, inverse modeling, model fusion, and integrated interpretations in the exploration of geophysics. A fundamental problem geophysical interpretation is proper geological understanding multiple inverted physical property images. Tackling this requires high-dimensional techniques for extracting information from modeled In study, we developed a 3D statistical tool extract features models based on synergy between independent component analysis continuous wavelet transform. An automated images also presented through hybrid spectral feature subset selection (SFSS) algorithm generalized supervised neural network rebuild limited targets Our self-proposed tested an Au/Ag epithermal system British Columbia (Canada), where layered volcano-sedimentary sequences, particularly felsic volcanic rocks, associated with mineralization. Geophysical were obtained cooperative inversion aeromagnetic, direct current resistivity, induced polarization sets. The recovered susceptibilities allowed locating magnetite destructive zone porphyritic intrusions volcanoes (Au host rocks). practical implementation SFSS study area shows that proposed learning scheme can efficiently learn lithotypes Au grade patterns makes predictions inputs. minimizes number extracted tries pick best representative each target case. This approach allows interpreters understand relevant irrelevant addition predictive models. Compared conventional interpolation methods, lithology add value deposit places without access prior borehole information.