Feature selection: We've barely scratched the surface Academic Article uri icon

abstract

  • Various feature selection techniques were studied. The trend toward richer data types is pushing feature selection in both scale and complexity. Natural language text features and image features are becoming commonplace. Several trends will increase the demand for feature selection. Time plays an additional role in some nonstationary domains where the best features have a seasonal dependency.

published proceedings

  • IEEE Intelligent Systems

author list (cited authors)

  • Forman, G., Liu, H., Dougherty, E. R., Dy, J. G., Torkkola, K., Tuv, E., ... Zhao, Z.

complete list of authors

  • Forman, G||Liu, H||Dougherty, ER||Dy, JG||Torkkola, K||Tuv, E||Peng, H||Ding, C||Long, F||Berens, M||Parsons, L||Yu, L||Zhao, Z

publication date

  • November 2005