Dataset binarization

Hi everybody,
is it possible with KNIME to “binarize” a dataset, like with WEKA filters.unsupervised.attribute.NominalToBinary? Any suggestions?

Thank you for your help
Cheers
Carmelo

I found the One2Many node… it should be the solution for doing that.

However, its output are integer values (0,1), but I need “nominal” values {f,t}. I used the NumberToString node, for converting them to string values, but passing this resulting data set to the Weka NaiveBayes classifier node, it says that it cannot handle string attributes…

Have you got any suggestions?

Thank you again
Cheers
Carmelo

Have you tried the native KNIME Naive Bayes node (Mining/Bayes/Naive Bayes Learner)? It works with String columns…

Ok, the native KNIME Naive Bayes node works, thank you. But it does not append prediction probabilities… is there any way for adding them?

I need: FalseProbablity, TrueProbability, Prediction

In the dialog of the Naive Bayes Predictor node you can set the checkbox to append the prediction probabilities (as described in the Node Description).