Duplicate: How to represent text for classification in weka?
What will be the input to the classifier if my data is Email, i.e, textual data? What will be the attributes/feature set in Email that I need to consider to perform classification based on discussion topic?