Academic Tweet Classification with Spreading activation based Label propagation algorithm using Tweet centric features
G. Manju, T. V. Geetha · 2016
Social network like Twitter is used by researchers and academicians to develop their professional relationship and as well it acts as a communication tool to share their research ideas, and research results. Among the enormous number of tweets, certain tweets contain general information and others might contain academic related information that aid scholarly communication. In order to identify such tweets, our work aims to classify the tweets as academic or non-academic tweet based on new combination of tweet content features like hashtags, terms that influence academic domain and user related features like follower/followee graph, mentions and tweet reply status. Spreading activation based Label propagation algorithm is used for tweet classification. In order to choose the appropriate random neighbor and effectively propagate the labels, the activation value computed by Spreading activation is used. Further, this helps to reduce the number of iterations and thereby, speeds up the convergence of the Label propagation algorithm. The comprehensive experimental results shows that spreading activation based Label propagation algorithm effectively classifies the tweets as academic or non-academic using the inferred academic feature combination with marginal decrease in the number of iterations.