Discovering temporal user interest on twitter by using semantic based dynamic interest finding model (TUT)

Social media systems are very popular in today's dynamic web. One of the famous social media systems is Twitter, in which peoples used to share their personal ideas about current issues with their friends. This work focuses on the problem of discovering a user's interest over time on twitter. Previous approaches have used to model the user topic of interest on twitter by building the profile of the users, that contain the words which can be used the user in his or her conversions with other users, but on twitter users used the noisy words which does not represent the correct topics or topic related to interest. This model has extended by a novel framework by using twitter user model. This model uses the latent topic variable to indicate the relatedness of the topic with any user. In this work, we propose a Temporal User Topic (TUT) approach which can consider the text of tweet by any user and time of the tweet. The proposed approach is used to discover topically related Users for different time periods. We also show how the interests and relationships of these users are changeovers a time period.