we recently found an issue with the Random Forest analysis. Specifically, each time we click the “Submit” button, the tool shows different sets of genes and outcome predictions. This inconsistency raises concerns about the reproducibility of the results.
For random forest, this means that many features in your data can give similar performance. Note this is true in general for omics data, reflecting statistical / biological / functional redundancies. You can turn off this by using “Randomness setting” (see below).
