Skip to main content
. Author manuscript; available in PMC: 2018 Oct 1.
Published in final edited form as: J Biomed Inform. 2017 Aug 9;74:1–9. doi: 10.1016/j.jbi.2017.08.005

RF average variable importance score

Set the number B of bootstrap samples
Step 1. For k = 1 to B
Draw a bootstrap sample from the data, run RF and generate the ranking of variable importance.
Let ac.reli,k be the importance assigned to the ith immune predictor in the kth RF run.
Step 2. For each immune predictor, compute its average score ac.reli=k=1Bac.reli,kB
Step 3. Sort the ac.relis to generate the ranking of average importance scores