A Feature Selection Method that Controls the False Discovery Rate

Mehdi Rostami, Olli Saarela

The problem of selecting a handful of truly relevant variables in supervised machine learning algorithms is a challenging problem in terms of untestable assumptions that must hold…