I have a result of text clustering based on TF-IDF. I have $k$ clusters. How can I get the representative terms for each cluster $I=1,\dots,k$ using the TF-IDF matrix?

Is there any standard way to do so?

Some preliminary ideas:

  • Calculate the average TF-IDF per term, take the highest values. This is equivalent to ranking the terms by sum of TF-IDF.
  • Fuzzy logic inspiration (fuzzy AND): Calculate the $f$ per term, take the highest values. Where $f$ can be minimum or product.
  • Concatenate all terms in all documents per cluster. Then perform TF-IDF on the merged document and rank the terms by this new score.
  • $\begingroup$ Hi @Karel Macek Did you solve your problem? I'm facing the same problem and finding solution. $\endgroup$ – thangdc94 Mar 18 at 2:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.