Detecting controversy on the web

Abstract

A useful feature to facilitate critical literacy would alert users when they are reading a controversial web page. This requires solving a binary classification problem: does a given web page discuss a controversial topic? We explore the feasibility of solving the problem by treating it as supervised k-nearest-neighbor classification. Our approach (1) maps a webpage to a set of neighboring Wikipedia articles which were labeled on a controversiality metric; (2) coalesces those labels into an estimate of the webpage's controversiality; and finally (3) converts the estimate to a binary value using a threshold. We demonstrate the applicability of our approach by validating it on a set of webpages drawn from seed queries. We show absolute gains of 22% in F0.5 on our test set over a sentiment-based approach, highlighting that detecting controversy is more complex than simply detecting opinions.

Content Bias in Online Health Search

“…Kacimi and Gamper's also present a summary of different arguments and sentiments related to the query topic. Other research has focused on the detection of controversy or disputed claims (including those with a medical focus) in Web pages [Ennals et al 2010;Dori-Hacohen and Allan 2013], the generation of summaries of opinions on political controversies [Awadallah et al 2012], and educating searchers about controversial topics by highlighting contrasting viewpoints [Vydiswaran et al 2012].…”

Something’s Brewing! Early Prediction of Controversy-causing Posts from Discussion Features

“…No, whether a topic (or entity/hashtag/word) has been controversial [a distinction also made by Addawood et al (2017)] (Popescu and Pennacchiotti, 2010;Choi et al, 2010;Cao et al, 2015;Lourentzou et al, 2015;Addawood et al, 2017;Al-Ayyoub et al, 2017;Garimella et al, 2018) No, whether a conversation contained disagreement (Mishne and Glance, 2006;Yin et al, 2012;Allen et al, 2014;Wang and Cardie, 2014) or mapping the disagreements (Awadallah et al, 2012;Marres, 2015;Borra et al, 2015;Liu et al, 2018) No, the task is, for the given textual item, predict antisocial behavior in the ensuing discussion (Zhang et al, 2018b,a), or subsequent comment volume/popularity/structure (Szabo and Huberman, 2010;Kim et al, 2011;Tatar et al, 2011;Backstrom et al, 2013;He et al, 2014;Zhang et al, 2018b), or eventual post article score (Rangwala and Jamali, 2010;Szabo and Huberman, 2010),; but all where, like us, the paradigm is early detection No, only info available at the item's creation (Dori-Hacohen and Allan, 2013;Mejova et al, 2014;Klenner et al, 2014;Addawood et al, 2017;Timmermans et al, 2017;Rethmeier et al, 2018;Kaplun et al, 2018) or the entire ensuing revision/discussion history (Rad and Barbosa, 2012;. N.B.…”

On the Retrieval of Wikipedia Articles Containing Claims on Controversial Topics

“…An additional line of works worth mentioning in the context of this work, are works on controversy detection, and more specifically works that detect controversial topics in Wikipedia [6,19]. While such works may be used to enhance existing manual controversy annotation in Wikipedia, focusing the retrieval solely on controversial Wikipedia articles provides a much less effective solution to the task, as shall be demonstrated in this work.…”