Mining latent relations in peer-production environments: a case study with Wikipedia article similarity and controversy

As people participate actively in social networking and peer-production sites, there are additional, implicit relations that emerge from various user activities. Mining such latent relations, or wisdom of crowds, is in itself an important area of ongoing research, with both general as well as domain-specific custom-made techniques. In this paper, we propose a new similarity measure, which we call expert-based similarity to discover semantic relations among Wikipedia articles from the co-editorship perspective. Also, different kinds of relations among entities may reveal diverse information. Both to explore and expose such a premise, we carry out a case study leveraging on multiple relations among Wikipedia articles. Specifically, we use expert-based similarity as well as other standard similarity measures, to discern the influence and impact of several factors which are hypothysed to generate controversies in Wikipedia articles. In the context of Wikipedia-specific research, our case study helps better differentiate the degree of impact of some of the possible causes of controversies.