Seminar: Mining Product Opinion from Tweets
Supervisor: Dr. Lourdes Pena-Castillo
"Mining Product Opinion from Tweets"
Department of Computer Science
Friday, December 11, 2015, 2:00 p.m., Room EN 2022
Social media networks facilitate people to communicate and express their opinion. The advent of Forums, Blogs and other social media has made people express opinions on any given topic freely thus making opinion mining essential not only for consumers to make a better purchase decision but also for the producers to make informed business decisions. This project focuses on an effective opinion mining approach for twitter. This study is entirely for mining the opinions posted for product based features. This analysis currently focuses on mining tweets reviewing mobile phones, laptops and bikes. The strategy used in mining is feature based (Evaluative factor). Reviews are classified as either good, bad or neutral. This approach provides an average precision and recall rates of 96% and 92% respectively for tweets in English. When tweets in South Indian regional language Tamil are considered, the precision and recall rates are 93% and 82% respectively.