I am trying to tweak my model in a way, that my random forest focuses on keeping the false negative number as low as possible since these have a much higer impact on my result than the false positives. Somewhere I read about a rule-node but I’m not sure how to apply it in a way, that my random forest prefers more FP in the result, if the FN is up to 50 times lower. Thank you very much in advance !
You can use the Binary Classification Inspector to adjust the threshold and see how your confusion matrix (and associated metrics) change. Once you arrive at a reasonable threshold value, you can apply a threshold as you mention using the Rule Engine node.
Check out this discussion from @paolotamag and associated workflow: