Robust Mahalanobis Distance

Data visualizations displayed on a laptop screen
Photo by Swello on Unsplash
projects

This methods project studies Minimum Covariance Determinant as a robust version of Mahalanobis distance, with the goal of improving the detection and treatment of multivariate outliers.

What already exists

  • Completed simulations
  • An early manuscript
  • A technically advanced foundation

Opportunity

The principal remaining work is writing, interpretation, and positioning. It may suit a quantitatively oriented student interested in robust statistics, simulation studies, or open-source methods.

Interested in helping revive this project? Complete the SAGE Lab Collaboration Assessment Form.