Opinion mining (sentiment mining, opinion/sentiment extraction) is the area of research that attempts to make automatic systems to determine human opinion from text written in natural language.
For example, a sample review of a new movie may look like this:I found the movie to be quite boring. It's not that I am solely interested in action movies, but any script that makes "My Dinner with Andre" seem fast-paced needs to be re-worked.
While doing research in this area, I have collected and prepared a number of training review datasets from various sources. I thought it would be helpful to share my datasets for download, so others can play around with them and compare results. These datasets were used in a couple of papers I wrote with Larry Yaeger at Indiana University. If you find these datasets useful, then we ask that you please cite the first paper.
Building a General Purpose Cross-Domain Sentiment Mining Model (Whitehead and Yaeger, CSIE, 2009)Whitehead, M. and Yaeger, L. Sentiment Mining Using Ensemble Classification Models, (SCSS), Dec. 2008
Opinion Mining/Sentiment Mining Dataset Downloads
Digital camera reviews from Amazon.com
Summer camp reviews from CampRatingz.com
Physician reviews from RateMDs.com
Pharmaceutical drug reviews from DrugRatingz.com
Laptop reviews from Amazon.com
Lawyer reviews from LawyerRatingz.com
Music (CD) reviews from Amazon.com
Radio show reviews from RadioRatingz.com
TV show reviews from TVRatingz.com