Elastic-net regularization versus -regularization for linear inverse problems with quasi-sparse solutions
arXiv:1604.03364 · doi:10.1088/1361-6420/33/1/015004
Abstract
We consider the ill-posed operator equation with an injective and bounded linear operator mapping between and a Hilbert space , possessing the unique solution \linebreak . For the cases that sparsity is expected but often slightly violated in practice, we investigate in comparison with the -regularization the elastic-net regularization, where the penalty is a weighted superposition of the -norm and the -norm square, under the assumption that . There occur two positive parameters in this approach, the weight parameter and the regularization parameter as the multiplier of the whole penalty in the Tikhonov functional, whereas only one regularization parameter arises in -regularization. Based on the variational inequality approach for the description of the solution smoothness with respect to the forward operator and exploiting the method of approximate source conditions, we present some results to estimate the rate of convergence for the elastic-net regularization. The occurring rate function contains the rate of the decay for and the classical smoothness properties of as an element in .
arXiv comment16 pages
References in corpus (6)
- Sparse Regularization with Penalty Term
- Convergence rates and source conditions for Tikhonov regularization with sparsity constraints
- Elastic-Net Regularization: Error estimates and Active Set Methods
- Convergence rates in -regularization if the sparsity assumption fails
- On the interplay of basis smoothness and specific range conditions occurring in sparsity regularization
- A unified approach to convergence rates for -regularization and lacking sparsity