Opinion Bias Detection with Social Preference Learning in Social Data

Opinion Bias Detection with Social Preference Learning in Social Data

A-Rong Kwon (Chonbuk National University, Jeonju-si, Korea) and Kyung-Soon Lee (Chonbuk National University, Jeonju-si, Korea)
Copyright: © 2013 |Pages: 20
DOI: 10.4018/ijswis.2013070104


In this paper, the authors propose a novel bias detection method based on social preference learning for targets on competing topics such as “GalaxyTab vs. iPad” in Twitter. People tend to evaluate a topic by expressing their opinions towards the associated targets such as price and quality. To exploit characteristics of social data, targets are extracted by a modified HITS algorithm on a tripartite graph. The main contribution is that social preferences are learned with explicit sentiment, latent sentiment as social semantics, and lexical sentiment as contextual semantics on targets of the topic, and that the individual preference is considered together with social preferences for the bias detection of a tweet. Experimental results on Twitter collection show significant improvements over all baseline methods. The results indicate that the method deals with not only the lack of a sentiment lexicon but also social and contextual semantics on targets of social users.
Article Preview


Now, social media has become an important way to communicate, socialize with others, and engage in arguments and debates (Popescu, Pennacchiotti & Paranjpe, 2011) via social network service (SNS) such as Twitter and Facebook. Twitter is an efficient conduit of information around the world, where massive numbers of instant messages (i.e. tweets) are freely published every day (Lumezanu, Feamster & Klein, 2012; Wang et al., 2011). The revolution in social data1 signifies the shift in human communication patterns towards increased personal information sharing, and hence shapes social opinions.

Whenever people make a choice, they are naturally inclined to hear public opinions, especially when the decision involves consuming valuable resources, such as time and/or money. Opinionated social data such as product reviews are now widely used by customers for decision making process (Mukherjee, Liu & Glance, 2012). For example, people are curious about how others think about the competing products such as “Galaxy Tab vs. iPad” or “Nikon vs. Canon” and it will offer greater convenience for them if social opinions are analyzed from the massive numbers of tweets in social network (Wang et al., 2011).

Researches on opinion analysis have attracted a great deal of attention recently due to many practical applications. Analysis of opinions, also known as opinion mining, sentiment analysis and subjectivity analysis, deals with the computational treatment of opinion, sentiment and subjectivity in text (Pang, Lee & Vaithyanathan, 2002). There are some connections to affective computing, where the goals include enabling computers to recognize and express emotions (Picard, 1997; D’Mello & Calvo, 2013). A semantic approach for sentiment analysis (Saif, He & Alani, 2012) uses semantic features by measuring the correlation of semantic entities and concepts with negative or positive sentiment for a training model in Twitter datasets.

Social preferences are a type of preferences studied in behavioral and experimental economics and social psychology. Social preferences can be learned from individual preference from massive amounts of social data. The individual’s opinion is generally dynamic, constantly changing from influence of other social opinions (Gao, 2012). Mining social opinions can play a primary role in helping individual’s decision making process. Researches on social opinion analysis are very important to reflect social semantics and social contexts within social media platforms because of the overwhelming amount of social data.

In this paper, we propose a novel bias detection method based on social preferences of targets learned automatically from Twitter data for the competing topics. Here, a topic is defined as the subject of something to be evaluated such as a product (‘Galaxy Tab’ or ‘iPad’) or a company (‘Samsung’ or ‘Apple’). The objective of opinion bias detection is to classify a user’s stance in a tweet as topic1–biased (topic1’s stance), topic2-biased or neutral stance for the two competing topics, topic1 and topic2, in social debates in order to analyze social opinions. People tend to evaluate a product-based topic by its associated targets (features) such as price, screen size, battery, and so on. Here, a target is defined as a key aspect of a topic towards which users can express either their positive or negative sentiments. However, target terms to express opinions on a topic are not equally important in a document; rather, they have various levels of strength to represent sentiments, which can be learned from social data.

Complete Article List

Search this Journal:
Open Access Articles
Volume 16: 4 Issues (2020): 1 Released, 3 Forthcoming
Volume 15: 4 Issues (2019)
Volume 14: 4 Issues (2018)
Volume 13: 4 Issues (2017)
Volume 12: 4 Issues (2016)
Volume 11: 4 Issues (2015)
Volume 10: 4 Issues (2014)
Volume 9: 4 Issues (2013)
Volume 8: 4 Issues (2012)
Volume 7: 4 Issues (2011)
Volume 6: 4 Issues (2010)
Volume 5: 4 Issues (2009)
Volume 4: 4 Issues (2008)
Volume 3: 4 Issues (2007)
Volume 2: 4 Issues (2006)
Volume 1: 4 Issues (2005)
View Complete Journal Contents Listing