This project was undertaken as a part of our Major Project during 2014-2015. We wanted to determine the factors which determine the popularity of a multimedia content.
The multimedia plaforms we studied are :
- FLICKR - Studied low level features such as HOG, LBP; higher level features such as presence of certain objects and certain social cues. Used Support Vector Regression to predict the views.
- Twitter - Devised a real-time clustering algorithm and an algorithm to rank topics.
- Youtube - Studied the impact of youtube trailers had on the actual IMDB ratings of the movie. We considered views, likes, sentiment polarity of the Youtube Comments.
- Varun Hegde(1hvarun@gmail.com)
- Santosh MS(sanms1993@gmail.com)
- Pramati Kalwad(pramati@gmail.com)