Hello
I'm using "InformationGainCalculator" node provided by Palladian. The speed is so low. Is there any other faster node for calculating information gain?
Best
Malik
Hi Malik,
probably you are working on a dataset with a wide feature space, right? The more features/columns, the more time it needs to compute the information gain for every step. I just tried the palladien node with a dataset containing 1000 columns and it really takes some time to execute. Afaik there is no standard node computing and outputting the information gain for you.
What's your intention behind this? Do you want to use this for reduction of feature space or somewhat similar? If so, you should give the KNIME learning content a try: https://www.knime.com/nodeguide/analytics/preprocessing.
Hope that helps.
Cheers,
Marten