Accelerated iterative hard thresholding algorithm for l0 regularized regression problem

In this paper, we propose an accelerated iterative hard thresholding algorithm for solving the l 0 regularized box constrained regression problem. We substantiate that there exists a threshold, if the extrapolation coefficients are chosen below this threshold, the proposed algorithm is equivalent to...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of global optimization Ročník 76; číslo 4; s. 819 - 840
Hlavní autoři: Wu, Fan, Bian, Wei
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York Springer US 01.04.2020
Springer Nature B.V
Témata:
Algorithms
Computer Science
Convergence
Iterative methods
Lower bounds
Mathematics
Mathematics and Statistics
Operations Research/Decision Theory
Optimization
Real Functions
R-linear convergence
Sparse regression problem
Accelerated iterative hard thresholding algorithm
regularization
ISSN:0925-5001, 1573-2916
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:In this paper, we propose an accelerated iterative hard thresholding algorithm for solving the l 0 regularized box constrained regression problem. We substantiate that there exists a threshold, if the extrapolation coefficients are chosen below this threshold, the proposed algorithm is equivalent to the accelerated proximal gradient algorithm for solving a corresponding constrained convex problem after finite iterations. Under some proper conditions, we get that the sequence generated by the proposed algorithm is convergent to a local minimizer of the l 0 regularized problem, which satisfies a desired lower bound. Moreover, when the data fitting function satisfies the error bound condition, we prove that both the iterate sequence and the corresponding sequence of objective function values are R-linearly convergent. Finally, we use several numerical experiments to verify our theoretical results.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0925-5001
1573-2916
DOI:10.1007/s10898-019-00826-6