Study level

  • PhD
  • Master of Philosophy
  • Honours

Faculty/School

Faculty of Science

School of Mathematical Sciences

Topic status

We're looking for students to study this topic.

Research centre

Centre for Data Science

Supervisors

Associate Professor Helen Thompson
Position
Associate Professor in Statistics
Division / Faculty
Faculty of Science
Dr Gentry White
Position
Associate Professor in Data Science and Government Statistics Chair
Division / Faculty
Faculty of Science

External supervisors

  • Claire Clarke, ABS
  • Edwin Lu, ABS

Overview

Machine learning cluster methods are common classification methods, but methods for assessing performance are limited as are methods for explaining how they work.  Exploring methods for both assessing and explaining performance are the subject of this research with application to real-world contexts with the Australian Bureau of Statistics.

Research activities

Research activities could include but are not limited to :

  • Discuss and recommend performance metrics for cluster analysis. Identify performance metrics that are method-agnostic (i.e., are applicable to all clustering methods and can be used to compare clustering methods).
  • Discuss and recommend explainability methods for cluster analysis.
  • Recommend other criteria for assessing the quality of the results from cluster analyses.
  • Testing can be done on ABS microdata.
  • If we can find an application of cluster analysis in the ABS, it may be feasible to make it a part of the PhD.

Outcomes

Outcomes will be commiserate with the level of research and engagement with the ABS

Skills and experience

A background in mathematics, statistics, or data or computer science are recommended

Keywords

Contact

Contact the supervisor for more information.