Name
Title | ||
|---|---|---|
Conditional Gradient Algorithmsfor Rank-One Matrix Approximationswith a Sparsity Constraint |
Abstract | ||
|---|---|---|
AbstractThe sparsity constrained rank-one matrix approximation problem is a difficult mathematicaloptimization problem which arises in a wide array of useful applications in engineering, machinelearning, and statistics, and the design of algorithms for this problem has attracted intensiveresearch activities. We introduce an algorithmic framework, called ConGradU, that unifies a varietyof seemingly different algorithms that have been derived from disparate approaches, and that allows forderiving new schemes. Building on the old and well-known conditional gradient algorithm, ConGradU isa simplified version with unit step size that yields a generic algorithm which either is given by ananalytic formula or requires a very low computational complexity. Mathematical properties aresystematically developed and numerical experiments are given. |
Year | DOI | Venue |
|---|---|---|
2011 | 10.1137/110839072 | Periodicals |
Keywords | Field | DocType |
sparse principal component analysis,PCA,conditional gradient algorithms,sparse eigenvalue problems,matrix approximations | Mathematical optimization,Matrix (mathematics),Mathematical analysis,Algorithm,Mathematical properties,Genetic algorithm,Mathematics,Computational complexity theory | Journal |
Volume | Issue | ISSN |
55 | 1 | 0036-1445 |
Citations | PageRank | References |
24 | 1.45 | 15 |
Authors | ||
2 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ronny Luss | 1 | 102 | 10.30 |
| Marc Teboulle | 2 | 5075 | 331.34 |