Advanced machine learning techniques for ore body modelling

Geostatistical techniques offer a means of mathematically approximating the spatial patterns of geological parameters. The geostatistical interpolation and simulation methods are commonly used for modeling ore bodies. The accuracy of these models has significant impacts on the reliability of mine planning and design. The proposed research project aims to apply advanced machine learning methods, geared specifically to the drillhole data type, to better predict the spatial distribution of rock properties (e.g. ore grade and rock hardness) in metallic ore bodies. Both classical and quantum machine learning techniques will be used to develop the ore body models. The accuracy of the machine learning models will be compared to the geostatistical models and with the field data.

Faculty Supervisor:

Kamran Esmaeili

Student:

Partner:

StratumAI Inc

Discipline:

Engineering

Sector:

Mining

University:

University of Toronto

Program: