Recognizing Experts on Social Media: A Heuristics-Based Approach

Knowing who is an expert on social media is a challenging yet important task, especially in a world where misleading information is commonplace and where social media is an important information source for knowledge seekers. In this paper we investigate expertise heuristics by comparing features of experts versus non-experts in big data settings. We employ a large set of features to classify experts and nonexperts using data collected on two social media platform (Twitter and reddit). Our results show a good ability to predict who is an expert, especially using language-based features, validating that heuristics can be developed to differentiate experts from novices organically, based on social media use. Our results contribute to the development of expertise location and identification systems as well as our understanding on how experts present themselves on social media.