Leveraging topic feature for followee recommendation on Twitter network

Abstract

With the fast growth of the Twitter network, users are overwhelmed by the huge amount of information, which is shared via the follower/followee social network, to overcome this problem, finding like-minded users becomes a very important task. Thus, a system to assist users in such a task is recommended. In this paper, we propose a followee recommendation system by leveraging the topic feature, for topic modeling, and the follower/followee topology, searching for similar users to recommend, based on topic similarities. To show the effectiveness of our approach, we evaluate it using a dataset ingathered from the Twitter platform. The experiment results indicate that our model outperforms the lexical-based [reference?] approach and semantic-based approach [reference?], achieving a recall value of more than 23% on recommending 10 followees, proving that dealing with users’ topics of interest in microblogging websites content is more efficient than semantic and lexical features.