The data set is selected from Multi Domain Sentiment Dataset, which can be found here: https://www.cs.jhu.edu/~mdredze/datasets/sentiment/index2.html

All credit to the authors.

Reference: John Blitzer, Mark Dredze, Fernando Pereira. Biographies, Bollywood, Boom-boxes and Blenders: Domain Adaptation for Sentiment Classification. Association of Computational Linguistics (ACL), 2007

---

Senti Word net is used for classification task.

It is obtained from following link. All credit for the word net goes to the authors: http://sentiwordnet.isti.cnr.it/