No abstract

This paper aims at introducing a real-time flaming mechanism that takes into account the behavior and trends of a user on a social networking site. We introduce a concept of layers through which better flaming can be implemented using Naive Bayesian theorem. Using this principle we aim to flame messages for a user according to the different kind of relationship which he/she enjoys with the peoples he/she has befriended on the social networking sites like Facebook, Linkedin, etc.