I am interested in performance analysis of classification algorithms. For this, I have collected tweets on a specific topic, and saved as .csv file. The .csv file consists of only one column called tweets that consists of the original text of the tweets. But to feed this .csv file to a classification algorithm in Orange, there must be a class or target column in it. The class or target column will tell if a tweet is Positive, Negative or Neutral. As long as there is no target column, it cannot be fed to a classifier.
If there were only a few tweets, I would start reading tweets one by one and manually mark them as Positive, Negative or Neutral in the target column. But for a very large number tweets, it will be very tedious and cumbersome task to manually create a target column with class values.
Can someone please help me automate the creation of target column? May be there is already a widget for for this task in Orange or some other way.