Intro to Machine Learning-Cluster Analysis 1-Day Course

Cluster analysis is a widely applied in machine learning and statistical analysis method. It has long been used in geochemical analysis for mineralogy and soil sampling in geological and agricultural sciences.

Cluster analysis is a data analysis technique used to categorise and label data sets by identifying groups (clusters). It is particularly valuable for classifying large or complex data sets with multiple variables and for comparing new data to established categories. Due to the size and complexity of modern data, cluster analysis relies on computers, making manual analysis impractical. Historically using these methods required programming expertise or specialist commercial software, limiting general usage. The availability of open-source tools like R and Python has significantly lowered the barriers to learning and applying these techniques.

The course contains Python based coding and requires very basic coding skills. The exercises are designed to be run by delegates with little to no coding skills. Once the software is loaded. The scripts are designed to be run in full without additional code being written. 

This course is aimed at Exploration Geologists, Mineral Resource Specialists. The course is set up to teach practical field methods and is not academically rigorous.   It is assumed that delegates have practical research or field experience. It is also assumed that delegates have a basic understanding of statistics and have some exposure to data analytics in geology, or geochemistry. 

Course content

Software, setup and Introduction to Python 

Software Setup and Introduction to Python and Jupyter Lab/Notebook – Instructions for loading the software before the course will be provided. A free Python intro course will also be provided for people who have never used Python. This takes about one afternoon to complete. 

Data used in this course will be multivariate data from base metals or bulk materials deposits and lithogeochemical data from public databases. Cluster analysis is intended for stratigraphic and lithological definition and, mapping of rock types and domaining. One of the use cases with be quality control and outlier identification.
 
This course looks at the two most popular methods for Cluster Analysis, Hierarchical Clustering and K-Means Clustering. We will cover the basic theory on the calculations and present several use cases including outlier detection, material types classification and stratigraphic identification in boreholes.

Computer and Software requirements 

• Materials
  • Laptop computer with administrative rights for installing the Anaconda and Python software -16GB RAM and a decent graphics card (gaming cards are ok)
  • Delegates will need to be able to access the venue WiFi internet for some parts of the course. 

• Software
  • Microsoft Excel (there are graphics and formulas that may not work as intended in Google Sheets)
  • Anaconda (Python) for Jupyter Notebook/Jupyter Lab (recommended for beginners or inexperienced coders) or VS_code(experience with coding in VS code required)
  • PDF reader 

Presenters

• Kathleen Body and Dennis Kattowitz from DEKA Dynamics.

Registration

Workshops must be purchased with a conference registration. If you wish to include the Machine Learning-Cluster Analysis workshop to your registration, please select the additional workshop option in your registration form.

If you have already registered to attend the Mineral Resource Estimation Conference and wish to add the workshop to your registration, please contact conference@ausimm.com.