CorpusBasedWeighing class_to_characterize

jaskiemr

I'm trying to do some text mining and I'm using the CorpusBasedWeighing. What does the class_to_characterize input parameter represent?
Thanks,
        mj

land

Hi,
this parameter defines, which class should be characterized by the weights. Please take a look in the operator doc for more information. The tool tip of the parameter will also help you a lot.

Greetings,
  Sebastian

jaskiemr

class to characterize: The target class for which to find characteristic feature weights.

So if I generate weights for the word vector, does the word list get labeled as a class, like "hurricanes" or something like that.

What can you do later with these classes?

Thank you for your help Sebastian,
          mj

land

Hi,
no, that's not what I meant. The word list just contains the words and their frequencies found in the documents of each class. And this isn't changed at all.
But there are attributes created from the word list. What you get is a weighting of these attributes. Just a weight per attribute. This weight should reflect how important this word is for  discriminating this class from the other classes.

Greetings,
  Sebastian