I received an inquiry —
“I’m doing some industry research, and I’d like to run some documents through a text analytics program to understand sentiment and key concepts/topics present in the documents. Probably a ‘few hundred’ documents over all.
“Can you recommend something simple and low cost that I could try? Perhaps something in the open source community?”
Here’s my response —
“I’m guessing you want to do minimal or no training, that is, you want a tool that will discover concepts and topics on its own? You could try Wordstat from Provalis Research (http://www.provalisresearch.com/wordstat/Wordstat.html) or Leximancer (https://leximancer.com/). Neither does sentiment out of the box although in principle, sentiment is just a classification problem that either would be able to handle. If you’re willing to do some training: Provalis has a tool called QDA Miner that supports coding and is linked to Wordstat.
“Your best bet, however, may be RapidMiner, which is free, open source. See http://rapid-i.com/content/view/184/196/ .
“I hope this helps. Please let me know what tool you chose and how it worked out.”
What (else) would YOU recommend?