Abstract

An introductory workshop designed for professionals working in all stages of the mine project life cycle.

The participants will be introduced to machine learning methods that can be applied to exploration, mine to mill optimization, and environmental planning. Practical demonstrations using Python and/or Orange will be performed using mostly geochemical datasets.

Morning session: Sam Wright

Data management and machine learning concepts will be introduced at a high-level and framed within the context of the mine project life cycle. The workshop will highlight techniques that help improve accuracy and efficiency of algorithms and/or workflows including best practices. Practical exercises will use either Python and open-source software such as Orange.

Topics covered:

• Introduction to Statistics – emphasis on compositional datasets

• Introduction to Machine learning – Supervised and Unsupervised

• Principal Component Analysis, with emphasis on biplot interpretation

• Decision Trees

• Algorithms - Clustering/Regression/Classification

• Model evaluations – Feature engineering, model metrics

Afternoon session: Tom Meuzelaar, Pim van Geffen, Juan Carlos Ordonez Calderon

Practical application of machine learning across the mine project life cycle will be featured. Case studies will emphasize methodology, pitfalls to consider, challenges and successes of employing advanced methods.

The session will end with a panel Q&A with all the presenters of the short course.

Short Course Objectives:

To provide participants with hands on examples of employing machine learning techniques across a mine project life cycle.

Target Audience:

Mining professionals working in all stages of the mine project life cycle and students.