I am using WEKA GUI to determine TF-IDF using unigrams and bigrams. I have used the "stringtowordvector filter) on my csv file and placed the following options to TRUE:

  • IDFTransform
  • TFTransform
  • OutputWordCounts and Tokenizer: N-gram tokenizer-max2-min1
However, after applying these to my data I don't know how to interpret/determine the TF-IDF value. I have assess to the table by clicking on the edit button:

I have tried and added the sum of each word and then order these sums but don't know if my approach is correct as I do not know what is the output or do I need to make any further calculations? Thank you in advance.