@article{fdi:010065259,
  title = {{S}parse zero-sum games as stable functional feature selection},
  author = {{S}okolovska, {N}. and {T}eytaud, {O}. and {R}izkalla, {S}. and {C}onsortium, {M}. and {C}lement, {K}. and {Z}ucker, {J}ean-{D}aniel},
  editor = {},
  language = {{ENG}},
  abstract = {{I}n large-scale systems biology applications, features are structured in hidden functional categories whose predictive power is identical. {F}eature selection, therefore, can lead not only to a problem with a reduced dimensionality, but also reveal some knowledge on functional classes of variables. {I}n this contribution, we propose a framework based on a sparse zero-sum game which performs a stable functional feature selection. {I}n particular, the approach is based on feature subsets ranking by a thresholding stochastic bandit. {W}e provide a theoretical analysis of the introduced algorithm. {W}e illustrate by experiments on both synthetic and real complex data that the proposed method is competitive from the predictive and stability viewpoints.},
  keywords = {},
  booktitle = {},
  journal = {{P}los {O}ne},
  volume = {10},
  numero = {9},
  pages = {e0134683 [16 p.]},
  ISSN = {1932-6203},
  year = {2015},
  DOI = {10.1371/journal.pone.0134683},
  URL = {https://www.documentation.ird.fr/hor/fdi:010065259},
}