This repository hosts an implementation of the convex-concave procedure (CCP) for robustly fitting a statistical factor model by maximum likelihood estimation. For more information, see our paper.
There are three notebooks:
- To recreate Figure 1, run PCA_outliers.ipynb.
- To recreate Figure 2, run MSE_run.ipynb.
- To recreate Figure 3, run DOA_run.ipynb.
If you wish to cite this work you may use the following BibTex:
@article{Cederberg24,
title = {Robust Factor Analysis Based on Tyler's M-estimator},
journal = {},
volume = {},
pages = {},
year = {},
issn = {},
doi = {},
author = {Daniel Cederberg},
}