Handling multi-label and Feature selection in classification

[Flake]

HI there, I wonder, how do you do multi-label classification (more specifically, text classification) in RM?

E.g. I'm thinking, having two labels for each item may improve the accuracy in classification given if the result from classification based on one label is used as feature in the other. However, this may be a wrong thinking, but anyway, does anyone know the best practice to use multi-label in classification? Any example?

Further, I wonder, given different classifier, say, NB, SVM, kNN, how do I view the features/dimensions used, and how can I manipulate the features used in these classifiers?

Thank you very much!

[pathros]

I have the same question:
I want to train a model with six different labels by means of a neural network.
i can only experiment with the following operator:
"Generate Multi-Label Data".
Nevertheless, the neural net asks me to define one label attribute!!!