Matches in SemOpenAlex for { <https://semopenalex.org/work/W3087730300> ?p ?o ?g. }
- W3087730300 abstract "We study the problem of predicting as well as the best linear predictor in a bounded Euclidean ball with respect to the squared loss. When only boundedness of the data generating distribution is assumed, we establish that the least squares estimator constrained to a bounded Euclidean ball does not attain the classical O(d∕n) excess risk rate, where d is the dimension of the covariates and n is the number of samples. In particular, we construct a bounded distribution such that the constrained least squares estimator incurs an excess risk of order Ω(d3∕2∕n) hence refuting a recent conjecture of Ohad Shamir [JMLR 2015]. In contrast, we observe that non-linear predictors can achieve the optimal rate O(d∕n) with no assumptions on the distribution of the covariates. We discuss additional distributional assumptions sufficient to guarantee an O(d∕n) excess risk rate for the least squares estimator. Among them are certain moment equivalence assumptions often used in the robust statistics literature. While such assumptions are central in the analysis of unbounded and heavy-tailed settings, our work indicates that in some cases, they also rule out unfavorable bounded distributions." @default.
- W3087730300 created "2020-09-25" @default.
- W3087730300 creator A5064204816 @default.
- W3087730300 creator A5071419221 @default.
- W3087730300 date "2023-02-01" @default.
- W3087730300 modified "2023-10-17" @default.
- W3087730300 title "Suboptimality of constrained least squares and improvements via non-linear predictors" @default.
- W3087730300 cites W1484867920 @default.
- W3087730300 cites W1506313179 @default.
- W3087730300 cites W1570963478 @default.
- W3087730300 cites W1575244755 @default.
- W3087730300 cites W1669551046 @default.
- W3087730300 cites W1967134148 @default.
- W3087730300 cites W1969705577 @default.
- W3087730300 cites W2003202757 @default.
- W3087730300 cites W2033484654 @default.
- W3087730300 cites W2043800867 @default.
- W3087730300 cites W2050159069 @default.
- W3087730300 cites W2054434031 @default.
- W3087730300 cites W2055734527 @default.
- W3087730300 cites W2079002396 @default.
- W3087730300 cites W2084572036 @default.
- W3087730300 cites W2103775046 @default.
- W3087730300 cites W2111396966 @default.
- W3087730300 cites W2129160848 @default.
- W3087730300 cites W2522906795 @default.
- W3087730300 cites W2954816895 @default.
- W3087730300 cites W2963283829 @default.
- W3087730300 cites W2964297554 @default.
- W3087730300 cites W2965064783 @default.
- W3087730300 cites W2995540461 @default.
- W3087730300 cites W3004992981 @default.
- W3087730300 cites W3036438388 @default.
- W3087730300 cites W3100353959 @default.
- W3087730300 cites W3100743579 @default.
- W3087730300 cites W3102654945 @default.
- W3087730300 cites W3103224927 @default.
- W3087730300 cites W3104823459 @default.
- W3087730300 cites W3105849782 @default.
- W3087730300 cites W3187217851 @default.
- W3087730300 cites W4205966363 @default.
- W3087730300 cites W4206039841 @default.
- W3087730300 cites W4211030719 @default.
- W3087730300 cites W4236362309 @default.
- W3087730300 cites W4249728363 @default.
- W3087730300 cites W4250954493 @default.
- W3087730300 cites W4375906137 @default.
- W3087730300 cites W648260396 @default.
- W3087730300 doi "https://doi.org/10.3150/22-bej1465" @default.
- W3087730300 hasPublicationYear "2023" @default.
- W3087730300 type Work @default.
- W3087730300 sameAs 3087730300 @default.
- W3087730300 citedByCount "3" @default.
- W3087730300 countsByYear W30877303002021 @default.
- W3087730300 crossrefType "journal-article" @default.
- W3087730300 hasAuthorship W3087730300A5064204816 @default.
- W3087730300 hasAuthorship W3087730300A5071419221 @default.
- W3087730300 hasBestOaLocation W30877303002 @default.
- W3087730300 hasConcept C105795698 @default.
- W3087730300 hasConcept C114614502 @default.
- W3087730300 hasConcept C118615104 @default.
- W3087730300 hasConcept C122041747 @default.
- W3087730300 hasConcept C129782007 @default.
- W3087730300 hasConcept C134306372 @default.
- W3087730300 hasConcept C185429906 @default.
- W3087730300 hasConcept C2524010 @default.
- W3087730300 hasConcept C2780069185 @default.
- W3087730300 hasConcept C28826006 @default.
- W3087730300 hasConcept C33676613 @default.
- W3087730300 hasConcept C33923547 @default.
- W3087730300 hasConcept C34388435 @default.
- W3087730300 hasConcept C9936470 @default.
- W3087730300 hasConceptScore W3087730300C105795698 @default.
- W3087730300 hasConceptScore W3087730300C114614502 @default.
- W3087730300 hasConceptScore W3087730300C118615104 @default.
- W3087730300 hasConceptScore W3087730300C122041747 @default.
- W3087730300 hasConceptScore W3087730300C129782007 @default.
- W3087730300 hasConceptScore W3087730300C134306372 @default.
- W3087730300 hasConceptScore W3087730300C185429906 @default.
- W3087730300 hasConceptScore W3087730300C2524010 @default.
- W3087730300 hasConceptScore W3087730300C2780069185 @default.
- W3087730300 hasConceptScore W3087730300C28826006 @default.
- W3087730300 hasConceptScore W3087730300C33676613 @default.
- W3087730300 hasConceptScore W3087730300C33923547 @default.
- W3087730300 hasConceptScore W3087730300C34388435 @default.
- W3087730300 hasConceptScore W3087730300C9936470 @default.
- W3087730300 hasIssue "1" @default.
- W3087730300 hasLocation W30877303001 @default.
- W3087730300 hasLocation W30877303002 @default.
- W3087730300 hasOpenAccess W3087730300 @default.
- W3087730300 hasPrimaryLocation W30877303001 @default.
- W3087730300 hasRelatedWork W1576985819 @default.
- W3087730300 hasRelatedWork W1989262972 @default.
- W3087730300 hasRelatedWork W2116799033 @default.
- W3087730300 hasRelatedWork W2143187693 @default.
- W3087730300 hasRelatedWork W2171635119 @default.
- W3087730300 hasRelatedWork W2467511638 @default.
- W3087730300 hasRelatedWork W2886701175 @default.
- W3087730300 hasRelatedWork W2921625644 @default.
- W3087730300 hasRelatedWork W4226060055 @default.