The increasing demand for nickel (Ni) and cobalt (Co), fueled by the clean energy transition, has heightened interest in Lachlan Orogen, Eastern Australia, for its lateritic Ni-Co resources. Despite recent discoveries, there remains a significant gap in understanding the extent of critical metals in the region. This study employs a machine learning-based framework that integrates multi-dimensional datasets to map the prospectivity of lateritic Ni-Co deposits in the Lachlan Orogen. It combines geological, structural, and geophysical data to create various models, from comprehensive to hybrid approaches focusing on different aspects of the mineral systems. The study also addresses machine learning challenges, such as data imbalance, using SMOTE-GAN and defines training samples effectively. The resulting prospectivity maps show a high correlation with known mineral occurrences, suggesting new areas for exploration and enhancing understanding of the mineralisation processes in the region.
- geopandas
- matplotlib
- netCDF4
- numpy
- pandas
- pulearn
- rioxarray
- skimage
- sklearn
- skopt
@article{Wake2024,
title = {Lateritic Ni-Co prospectivity modelling in eastern Australia using an enhanced generative adversarial network and positive-unlabelled bagging},
author = {Wake, N. and Farahbakhsh, Ehsan and M{\"u}ller, R. Dietmar},
year = {2024},
journal = {?},
volume = {?},
number = {?},
pages = {?},
doi = {?},
}