Geological domaining with unsupervised clustering and ensemble support vector classification

A geological model accounting for uncertainties possesses important advantages for resource estimation. Machine learning algorithms (MLAs) employed on multivariate geochemical datasets open up ways to new methodologies for such geological models with ease in comparison to traditional geostatistical methods. This article proposes a two-step MLA with an ensemble implementation to define geological domains and their uncertainties based on geochemical data. The proposed workflow is applied hierarchically on a dataset from a porphyry copper deposit to perform binary classification that can be attributed to alteration domains.