Loading...
Please wait, while we are loading the content...
Similar Documents
A class of operator splitting methods for least absolute shrinkage and selection operator (LASSO) models
| Content Provider | Semantic Scholar |
|---|---|
| Author | Mo, Lili |
| Copyright Year | 2013 |
| Abstract | Sparsity of statistical models is important to interpret high dimensional data. Linear regression models with sparsity have various applications in many real-world processes, including signal processing, image processing, gene selection and comparative genomic hybridization. Commonly, it involves heavy computation for combinatorial search for finding such models with sparsity by model selection. The least absolute shrinkage and selection operator (LASSO) is a popular technique for model selection and estimation in linear regression to induce sparsity. In this thesis, we study several general LASSO models, including group LASSO, sparse group LASSO and fused LASSO. We focus on the applications of alternating direction method (ADM) of multipliers to these LASSO models, including the applications of ADM to LASSO and group LASSO, and a linearized version of ADM to sparse group LASSO and fused LASSO. This approach can solve large scale problems efficiently. Some numerical study for colon cancer and leukemia cancer data sets are reported. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://lib-nt2.hkbu.edu.hk/cil-image/theses/abstracts/b37117853a.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
