Equal Size Sampling

I see that Knime has the " Equal Size Sampling" node that picks the equal size of each class.
Is there another Knime node that allows you to choose with different ratios? for example, if I have data with 1000 positive and 40 negatives then I want to pick 40 negatives and 80 positives or another amount of positives.


How about you split your data by positive and negative, randomly select the number of cases you want and concatenate them back together. You could set parameters for that to make it more flexible.

Yes that’s works- I thought there is a Knime node


1 Like