Concept Similarity Based Academic Tweet Community Detection Using Label Propagation

摘要

In today's world, Social Network plays a vital role in the society. Social Network users share their ideas, views, opinions, and develop their personal relationship. Social Network has major influence with academic community. This paper aims at detecting similar concept based academic tweets from the numerous available tweets and forming a community, considering the social relation between the tweeters. Academic community can support recommender system for researcher network. In our work, in order to extract concept similarity based academic community, concept similarity graph is constructed from twitter. Label Propagation algorithm is used to detect academic community. Normally, tweets contain user views, suggestions and discussion on a specific topic. In spite of tweets, containing other words in it, Concept words play a vital role in identifying about the aim of the tweeter in posting the tweet. Moreover, for academic topics, academic concepts are important. So, the Concepts are extracted and based on the similarity between concepts, academic community has been extracted from twitter. Label propagation has proven to be an effective method for detecting communities in complex networks. In this work, the new update rule based on social relation is introduced for Label propagation algorithm and used for concept based community detection. The experiment shows that, in comparison with standard label propagation algorithm, the label propagation with modified update rule reduces the number of iterations for convergence and as well was more effective in detecting communities.