Automated facies classification and metal grade prediction using machine learning algorithms
- Datum: –16.00
- Plats: Geocentrum Dk235
- Föreläsare: Glen Nwaila, School of Geosciences, University of the Witwatersrand
- Kontaktperson: Monika Ivandic
Machine learning is a broad subfield of artificial intelligence and refers to a set of tools for modelling and predicting patterns derived from complex datasets. In geosciences, machine learning algorithms form the basis of the fourth industrial revolution and have direct implications for future methodologies that will be employed for the exploration and exploitation of mineral deposits. Here we introduce a set of algorithms to predict metal grade and perform automated facies identification called GS-Pred. GS-Pred uses geological descriptions and metal assay data and was developed by Dr Glen Nwaila and Dr Steven Zhang (School of Geosciences, University of the Witwatersrand). Results from GS-Pred are compared with those from existing machine learning algorithms such as non-linear support vector regression (SVR) with a radial basis function kernel, non-parametric regression, K-nearest neighbours (KNN) and penalised linear regression, elastic net regression with cross-validation (“ElasticNetCV”). Current projects on this work include the application of artificial neural networks for target exploration.
Brief biographic sketch:
A geoscientist with multi-disciplinary experience in geology and process engineering currently employed as Lecturer of Economic Geology at the School of Geosciences, Wits University. Glen's formal education includes various qualifications that span BSc (Hons) Geology (University of Johannesburg, South Africa), MSc Chemical Engineering with merit (University of Cape Town, South African) and a PhD (Magna Cum Laude) Economic Geology (University of Würzburg, Germany).