RGIFE - Ranked Guided Iterative Feature Elimination

RGIFE is a feature reduction heuristic for the identification of small panels of highly predictive biomarkers. The heuristic is based on an iterative reduction paradigm: first train a classifier, then rank the attributes based on their importance and finally remove attributes in block. RGIFE is designed to work with large-scale datasets and identifies reduced set of attributes with high classification power.

Documentation

To learn how to run and configure RGIFE read the tutorial.

Download

You can download the RGIFE source code (GNU GPLv3).

Project team

• Nicola Lazzarini
• Jaume Bacardit

Publications

• DOI BibTeX
  N. Lazzarini, J. Bacardit
  RGIFE: a ranked guided iterative feature elimination heuristic for the identification of biomarkers
  in BMC Bioinformatics, 18(1), December 2017

  @ARTICLE{Lazzarini2017,
    title = {RGIFE: a ranked guided iterative feature elimination heuristic for the identification of biomarkers},
    author = {Lazzarini, Nicola and Bacardit, Jaume},
    year = 2017,
    doi = {10.1186/s12859-017-1729-2},
    month = dec,
    journal = {BMC Bioinformatics},
    volume = {18},
    number = {1}