Feature Selection in Weakly Coherent Matrices
| Affiliation auteurs | Affiliation ok |
| Titre | Feature Selection in Weakly Coherent Matrices |
| Type de publication | Conference Paper |
| Year of Publication | 2018 |
| Auteurs | Chretien S, Ho O |
| Editor | Deville Y, Gannot S, Mason R, Plumbley MD, Ward D |
| Conference Name | LATENT VARIABLE ANALYSIS AND SIGNAL SEPARATION (LVA/ICA 2018) |
| Publisher | Ctr Vis Speech & Signal Proc; Inst Sound Recording |
| Conference Location | GEWERBESTRASSE 11, CHAM, CH-6330, SWITZERLAND |
| ISBN Number | 978-3-319-93764-9; 978-3-319-93763-2 |
| Mots-clés | Coherence, Null space property, Restricted invertibility |
| Résumé | A problem of paramount importance in both pure (Restricted Invertibility problem) and applied mathematics (Feature extraction) is the one of selecting a submatrix of a given matrix, such that this submatrix has its smallest singular value above a specified level. Such problems can be addressed using perturbation analysis. In this paper, we propose a perturbation bound for the smallest singular value of a given matrix after appending a column, under the assumption that its initial coherence is not large, and we use this bound to derive a fast algorithm for feature extraction. |
| DOI | 10.1007/978-3-319-93764-9_13 |