Cyber Times International Journal of Technology & Management

Twitter is well-known Online Social Networking started in March 2006.It first became widespread among high profile people like politicians, celebrities and then millions of their fans started joining twitter and following their favourite celebrity. Fans get up-to-date information posted by their favourite celebrities, and also comment, retweet and like their posts. But some of the notorious people take advantage of this and misuse it as they speculate about celebrities’ posts and spread wrong rumours by writing blogs and anonymously commenting on twitter and like this they breach their privacy. In this paper sentiment analysis technique is applied on the posts containing  the offensive language which has been collected by Twitter Streaming API and then the collected tweets are classified by Naïve Bayes classifier and Maximum Entropy. The accuracy given by Naïve Bayes classifier is approx. 81.2% and Maximum Entropy classifier is approx. 79.7%.