Evaluating the impact of topological protein features on the negative examples selection (BMC Bioinformatics)

Supervised machine learning methods when applied to the problem of automated protein-function prediction (AFP) require the availability of both positive examples (i.e., proteins which are known to possess a given...