Shrinkage MMSE estimators of covariances beyond the zero-mean and stationary variance assumptions
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913405966745600 |
|---|---|
| author | Flasseur, Olivier Thiébaut, Eric Denis, Loïc Langlois, Maud |
| author_facet | Flasseur, Olivier Thiébaut, Eric Denis, Loïc Langlois, Maud |
| contents | We tackle covariance estimation in low-sample scenarios, employing a structured covariance matrix with shrinkage methods. These involve convexly combining a low-bias/high-variance empirical estimate with a biased regularization estimator, striking a bias-variance trade-off. Literature provides optimal settings of the regularization amount through risk minimization between the true covariance and its shrunk counterpart. Such estimators were derived for zero-mean statistics with i.i.d. diagonal regularization matrices accounting for the average sample variance solely. We extend these results to regularization matrices accounting for the sample variances both for centered and non-centered samples. In the latter case, the empirical estimate of the true mean is incorporated into our shrinkage estimators. Introducing confidence weights into the statistics also enhance estimator robustness against outliers. We compare our estimators to other shrinkage methods both on numerical simulations and on real data to solve a detection problem in astronomy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_07104 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Shrinkage MMSE estimators of covariances beyond the zero-mean and stationary variance assumptions Flasseur, Olivier Thiébaut, Eric Denis, Loïc Langlois, Maud Instrumentation and Methods for Astrophysics Methodology We tackle covariance estimation in low-sample scenarios, employing a structured covariance matrix with shrinkage methods. These involve convexly combining a low-bias/high-variance empirical estimate with a biased regularization estimator, striking a bias-variance trade-off. Literature provides optimal settings of the regularization amount through risk minimization between the true covariance and its shrunk counterpart. Such estimators were derived for zero-mean statistics with i.i.d. diagonal regularization matrices accounting for the average sample variance solely. We extend these results to regularization matrices accounting for the sample variances both for centered and non-centered samples. In the latter case, the empirical estimate of the true mean is incorporated into our shrinkage estimators. Introducing confidence weights into the statistics also enhance estimator robustness against outliers. We compare our estimators to other shrinkage methods both on numerical simulations and on real data to solve a detection problem in astronomy. |
| title | Shrinkage MMSE estimators of covariances beyond the zero-mean and stationary variance assumptions |
| topic | Instrumentation and Methods for Astrophysics Methodology |
| url | https://arxiv.org/abs/2403.07104 |