File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1214/18-AOS1759
- Scopus: eid_2-s2.0-85078981162
- WOS: WOS:000493896800003
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: Sorted concave penalized regression
Title | Sorted concave penalized regression |
---|---|
Authors | |
Keywords | Concave penalties Local convex approximation Minimax rate Penalized least squares Restricted eigenvalue Signal strength Slope Sorted penalties |
Issue Date | 2019 |
Citation | Annals of Statistics, 2019, v. 47, n. 6, p. 3069-3098 How to Cite? |
Abstract | The Lasso is biased. Concave penalized least squares estimation (PLSE) takes advantage of signal strength to reduce this bias, leading to sharper error bounds in prediction, coefficient estimation and variable selection. For prediction and estimation, the bias of the Lasso can be also reduced by taking a smaller penalty level than what selection consistency requires, but such smaller penalty level depends on the sparsity of the true coefficient vector. The sorted ℓ1 penalized estimation (Slope) was proposed for adaptation to such smaller penalty levels. However, the advantages of concave PLSE and Slope do not subsume each other. We propose sorted concave penalized estimation to combine the advantages of concave and sorted penalizations. We prove that sorted concave penalties adaptively choose the smaller penalty level and at the same time benefits from signal strength, especially when a significant proportion of signals are stronger than the corresponding adaptively selected penalty levels. A local convex approximation for sorted concave penalties, which extends the local linear and quadratic approximations for separable concave penalties, is developed to facilitate the computation of sorted concave PLSE and proven to possess desired prediction and estimation error bounds. Our analysis of prediction and estimation errors requires the restricted eigenvalue condition on the design, not beyond, and provides selection consistency under a required minimum signal strength condition in addition. Thus, our results also sharpens existing results on concave PLSE by removing the upper sparse eigenvalue component of the sparse Riesz condition. |
Persistent Identifier | http://hdl.handle.net/10722/318812 |
ISSN | 2023 Impact Factor: 3.2 2023 SCImago Journal Rankings: 5.335 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Feng, Long | - |
dc.contributor.author | Zhang, Cun Hui | - |
dc.date.accessioned | 2022-10-11T12:24:37Z | - |
dc.date.available | 2022-10-11T12:24:37Z | - |
dc.date.issued | 2019 | - |
dc.identifier.citation | Annals of Statistics, 2019, v. 47, n. 6, p. 3069-3098 | - |
dc.identifier.issn | 0090-5364 | - |
dc.identifier.uri | http://hdl.handle.net/10722/318812 | - |
dc.description.abstract | The Lasso is biased. Concave penalized least squares estimation (PLSE) takes advantage of signal strength to reduce this bias, leading to sharper error bounds in prediction, coefficient estimation and variable selection. For prediction and estimation, the bias of the Lasso can be also reduced by taking a smaller penalty level than what selection consistency requires, but such smaller penalty level depends on the sparsity of the true coefficient vector. The sorted ℓ1 penalized estimation (Slope) was proposed for adaptation to such smaller penalty levels. However, the advantages of concave PLSE and Slope do not subsume each other. We propose sorted concave penalized estimation to combine the advantages of concave and sorted penalizations. We prove that sorted concave penalties adaptively choose the smaller penalty level and at the same time benefits from signal strength, especially when a significant proportion of signals are stronger than the corresponding adaptively selected penalty levels. A local convex approximation for sorted concave penalties, which extends the local linear and quadratic approximations for separable concave penalties, is developed to facilitate the computation of sorted concave PLSE and proven to possess desired prediction and estimation error bounds. Our analysis of prediction and estimation errors requires the restricted eigenvalue condition on the design, not beyond, and provides selection consistency under a required minimum signal strength condition in addition. Thus, our results also sharpens existing results on concave PLSE by removing the upper sparse eigenvalue component of the sparse Riesz condition. | - |
dc.language | eng | - |
dc.relation.ispartof | Annals of Statistics | - |
dc.subject | Concave penalties | - |
dc.subject | Local convex approximation | - |
dc.subject | Minimax rate | - |
dc.subject | Penalized least squares | - |
dc.subject | Restricted eigenvalue | - |
dc.subject | Signal strength | - |
dc.subject | Slope | - |
dc.subject | Sorted penalties | - |
dc.title | Sorted concave penalized regression | - |
dc.type | Article | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1214/18-AOS1759 | - |
dc.identifier.scopus | eid_2-s2.0-85078981162 | - |
dc.identifier.volume | 47 | - |
dc.identifier.issue | 6 | - |
dc.identifier.spage | 3069 | - |
dc.identifier.epage | 3098 | - |
dc.identifier.eissn | 2168-8966 | - |
dc.identifier.isi | WOS:000493896800003 | - |