Most informative feature sets for breast cancer survival prediction in METABRIC data.

AKLIMATE weights were averaged over 50 train/test splits. The table lists the 20 most relevant feature sets, out of 1836 feature sets with a non-zero weight in at least one train/test split. Weights were normalized to sum to 1. Aliases for the top 10 feature sets are included in brackets—they provide a more descriptive name for the underlying biological function, based on information gathered from the source publication. The aliases are used in Fig 2 and throughout the text.

(PDF)