Bouguessa, Mohamed, and Lotfi Ben Romdhane. "Identifying authorities in online communities." ACM Transactions on Intelligent Systems and Technology (TIST) 6.3 (2015): 1-23.

This article presents a novel parameterless mixture model-based approach to address the challenges of identifying authoritative users in online communities. Traditional methods face issues like the absence of an automatic mechanism for distinguishing authoritative users, reliance on supervised learning which depends heavily on scarce labeled data, and the inaccuracy caused by using incorrect user parameters. The proposed solution utilizes a feature vector for each user, reflecting their social behavior and activity, and employs a statistical framework based on multivariate beta mixtures to model these vectors. The approach effectively identifies the most authoritative users by estimating the probability density function and pinpointing the beta component associated with authority. This methodology's effectiveness is demonstrated using real data from Stack Exchange and Twitter.