Keyword extraction from micro-blogs using collective weight

Abstract

The growth of social networking has increased the scope of expression on a public platform. Twitter alone, being one of the most trending social networking sites, generates a huge amount of text every minute. Twitter content analysis and summarization benefits many applications such as information retrieval, automatic indexing, automatic classification, automatic clustering, automatic filtering, etc. One of the most important tasks in analyzing tweets is automatic keyword extraction. Many existing graph-based keyword extraction approaches determine keywords purely based on centrality measure. However, various features such as frequency, centrality, position, and strength of the neighbours of the keyword also affect the importance of a keyword in tweets. Therefore, this paper proposes a novel unsupervised graph-based keyword extraction method called keywords from collective weights (KCW) which determines the importance of a keyword by collectively considering various influencing features. The KCW is based on node-edge rank centrality with node weight depending on various features. The model is validated with five data sets: Uri Attack, Harry Potter, IPL, Donald Trump and IPhone5. The result of KCW is compared with three existing models. It is observed from the experimental results that the proposed method is far better than the others. The performances are shown in terms of precision, recall, and F measure.