Hello,

Have a look at this;

http://www.cs.jhu.edu/~mdredze/datasets/sentiment/ for training purposes. I have used a dataset having item reviews rated from 1–5 stars. I downloaded processed_stars.tar.gz file, extracted and gathered all all_balanced.review files in subfolders to a separate path data/processed_stars_all.

This is how the data/processed_stars_all comes. First, you have to download the data and check for files. I use the star ratings for the analysis.