Resource Management Through Machine Learning

In the modern era of diminishing returns on fixed exploration budgets, challenging targets, and ever-increasing numbers of multi-parameter datasets, proper management and integration of available data is a crucial component of any resource exploration program. Machine learning algorithms have successfully been used for years by the technology sector to accomplish just this task on their databases, and recent developments aim at appropriating these successes to the field of natural resource exploration. Numerous algorithms have been attempted for resource prospectivity mapping in the past, and in this paper we apply a modified support-vector machine algorithm to a test dataset from the QUEST region in central British Columbia, Canada, to target undiscovered Cu-Au porphyry districts. The modified algorithm is designed to properly handle the highly variable uncertainty associated with both the training data (ie: geophysics, geochemistry, geological mapping) as well as the training labels (known Cu-Au porphyry targets in the region). Support vector machines are introduced, the challenges of working with geoscientific datasets are discussed, and finally results from applying the modified algorithm to the QUEST dataset are presented.