Is Wikipedia Really Neutral? A Sentiment Perspective Study of War-related Wikipedia Articles since 1945

Is Wikipedia Really Neutral? A Sentiment Perspective Study of War-related Wikipedia Articles since 1945 Yiwei Zhou, Alexandra I. Cristea and Zachary Roberts Department of Computer Science University of Warwick Coventry, United Kingdom {Yiwei.Zhou, A.I.Cristea, Z.L.Roberts}@warwick.ac.uk Abstract Wikipedia is supposed to be supporting the “Neutral Point of View”. Instead of accept- ing this statement as a fact, the current paper analyses its veracity by specifically analysing a typically controversial (negative) topic, such as war, and answering questions such as “Are there sentiment differences in how Wikipedia articles in different languages describe the same war?”. This paper tackles this chal- lenge by proposing an automatic methodology based on article level and concept level senti- ment analysis on multilingual Wikipedia arti- cles. The results obtained so far show that rea- sons such as people’s feelings of involvement and empathy can lead to sentiment expression differences across multilingual Wikipedia on war-related topics; the more people contribute to an article on a war-related topic, the more extreme sentiment the article will express; dif- ferent cultures also focus on different concepts about the same war and present different sen- timents towards them. Moreover, our research provides a framework for performing differ- ent levels of sentiment analysis on multilin- gual texts. 1 Introduction Wikipedia is the largest and most widely used encyclopaedia in collaborative knowledge building (Medelyan et al., 2009). Since its start in 2001, it contains more than 33 million articles in more than 200 languages, while only about 4 million articles are in English1. Possible sources for the content 1http://meta.wikimedia.org/wiki/List_of_ Wikipedias include books, journal articles, newspapers, web- pages, sound recordings2, etc. Although a “Neutral point of view” (NPOV)3 is Wikipedia’s core content policy, we believe sentiment expression is inevitable in this user-generated content. Already in (Green- stein and Zhu, 2012), researchers have raised doubt about Wikipedia’s neutrality, as they pointed out that “Wikipedia achieves something akin to a NPOV across articles, but not necessarily within them”. Moreover, people of different language backgrounds share different cultures and sources of information. These differences have reflected on the style of con- tributions (Pfeil et al., 2006) and the type of informa- tion covered (Callahan and Herring, 2011). Further- more, Wikipedia webpages actually allow to con- tain opinions, as long as they come from reliable au- thors4. Due to its openness to multiple forms of con- tribution, the articles on Wikipedia can be viewed as a summarisation of thoughts in multiple languages about specific topics. Automatically detecting and measuring the differences can be crucial in many applications: public relation departments can get some useful suggestions from Wikipedia about top- ics close to their hearts; Wikipedia readers can get some insights about what people speaking other lan- guages think about the same topic; Wikipedia ad- ministrators can quickly locate the Wikipedia arti- cles that express extreme sentiment, to better apply the NPOV policy, by eliminating some edits. 2http://en.wikipedia.org/wiki/Wikipedia: Citing_sources 3http://en.wikipedia.org/wiki/Wikipedia: Neutral_point_of_view 4http://en.wikipedia.org/wiki/Wikipedia: Identifying_reliable_sources PACLIC 29 160 29th Pacific Asia Conference on Language, Information and Computation pages 160 - 168 Shanghai, …

Content truncated to 500 words