Twitterrank: Finding Topic-Sensitive Influential Twitterers

This article focuses on the problem of identifying influential Twitter users using Machine Learning (ML) techniques and Natural Language Processing (NLP). The authors argue that measuring followers is not an effective means to determine influence; that there is generally a high degree of reciprocal following relationships on Twitter that can be cordial and meaningless. The authors analyze a set of top-1000 Singapore-base Twitterers to assess their approach. The authors propose that TwitterRank, an extension of the PageRank algorithm, can be used effectively to measure the influence of individual users on Twitter. In their study, the authors construct topic-sensitive relationship networks, and apply the TwitterRank algorithm to measure influence. Their process takes both the topical similarity between users and the link structure into account.

This article will be of use to researchers and practitioners interested in identifying influential users within micro-blogging and social media platforms such as Twitter. According to the authors, the presence of reciprocity can be explained by the phenomenon of homophily, in which users with similar interests, sociodemographic, behavioural and intrapersonal characteristics, will follow one another. The authors find that the TwitterRank algorithm outperforms other related algorithms in identifying topic-sensitive influential Twitterers, and a number of future steps are provided as a means to guide further research into the topic.