Academic CS in Production Code

Influence Function and Efficiency of the Minimum Covariance Determinant Scatter Matrix Estimator

C. Croux, G. Haesbroeck
1999
6 references

Abstract

The minimum covariance determinant (MCD) scatter estimator is a highly robust estimator for the dispersion matrix of a multivariate, elliptically symmetric distribution. It is relatively fast to compute and intuitively appealing. In this note we derive its influence function and compute the asymptotic variances of its elements. A comparison with the one step reweighted MCD and with S-estimators is made. Also finite-sample results are reported.

View Paper PDF DOI
🤖 Machine Learning
1 repository
6 references

Code References

â–¶ scikit-learn/scikit-learn
1 file
â–¶ sklearn/covariance/_robust_covariance.py
6
.. [Croux1999] Croux, C., Haesbroeck, G. "Influence Function and
Correction using the asymptotic correction factor derived by [Croux1999]_.
.. [Croux1999] Influence Function and Efficiency of the Minimum
distribution, following [Croux1999]_.
.. [Croux1999] Influence Function and Efficiency of the Minimum
# Parameter alpha as in [Croux1999] Eq. 4.2
Link copied to clipboard!