A corpus study of evaluative and speculative language

Wiebe, Janyce; Bruce, Rebecca; Bell, Matthew; Martin, Melanie; Wilson, Theresa

doi:10.3115/1118078.1118104

Cited by 31 publications

(28 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Rubin (2006b) traces the roots of subjectivity identification tools to the work of (Wiebe et al, 2001) who proposed one of the first annotation schemes to classify and identify subjective and objective statements in texts. Prior to this work on subjectivity, Rubin (2006b) continues, an NLP system needed to determine the structure of a text -normally at least enough to answer "Who did what to whom?"…”

Section: Subjectivity and Opinion Mining Or Sentiment Analysismentioning

confidence: 99%

Deception Detection and Rumor Debunking for Social Media

Rubin¹

2016

The SAGE Handbook of Social Media Research Methods

View full text Add to dashboard Cite

The main premise of this chapter is that the time is ripe for more extensive research and development of social media tools that filter out intentionally deceptive information such as deceptive memes, rumors and hoaxes, fake news or other fake posts, tweets and fraudulent profiles. Social media users' awareness of intentional manipulation of online content appears to be relatively low, while the reliance on unverified information (often obtained from strangers) is at an all-time high. I argue there is need for content verification, systematic fact-checking and filtering of social media streams. This literature survey provides a background for understanding current automated deception detection research, rumor debunking, and broader content verification methodologies, suggests a path towards hybrid technologies, and explains why the development and adoption of such tools might still be a significant challenge. Biographical DetailsVictoria L. Rubin is an Associate Professor at the Faculty of Information and Media Studies and the Director of the Language and Information Technologies Research Lab (LiT.RL) at the University of Western Ontario. She specializes in information retrieval and natural language processing techniques that enable analyses of texts to identify, extract, and organize structured knowledge. She studies complex human information behaviors that are, at least partly, expressed through language such as deception, uncertainty, credibility, and emotions. Her research on Deception Detection has been published in recent core workshops on the topic and prominent information science conferences, as well as the

show abstract

Section: Subjectivity and Opinion Mining Or Sentiment Analysismentioning

confidence: 99%

Deception Detection and Rumor Debunking for Social Media

Rubin¹

2016

The SAGE Handbook of Social Media Research Methods

View full text Add to dashboard Cite

show abstract

“…Rubin (2006) traces the roots of subjectivity/objectivity identification work in NLP to Wiebe, Bruce, Bell, Martin, and Wilson (2001) who developed one of the first annotation schemes to classify and identify subjective and objective statements in texts. Prior to this work on subjectivity, Rubin (2006) continues, an NLP system needed to determine the structure of a text -normally at least enough to answer "Who did what to whom?"…”

Section: Literature Reviewmentioning

confidence: 99%

Veracity Roadmap: Is Big Data Objective, Truthful and Credible?

Lukoianova

Rubin

2014

ACRO

138

View full text Add to dashboard Cite

This paper argues that big data can possess different characteristics, which affect its quality. Depending on its origin, data processing technologies, and methodologies used for data collection and scientific discoveries, big data can have biases, ambiguities, and inaccuracies which need to be identified and accounted for to reduce inference errors and improve the accuracy of generated insights. Big data veracity is now being recognized as a necessary property for its utilization, complementing the three previously established quality dimensions (volume, variety, and velocity), But there has been little discussion of the concept of veracity thus far. This paper provides a roadmap for theoretical and empirical definitions of veracity along with its practical implications. We explore veracity across three main dimensions: 1) objectivity/subjectivity, 2) truthfulness/deception, 3) credibility/implausibility -and propose to operationalize each of these dimensions with either existing computational tools or potential ones, relevant particularly to textual data analytics. We combine the measures of veracity dimensions into one composite index -the big data veracity index. This newly developed veracity index provides a useful way of assessing systematic variations in big data quality across datasets with textual information. The paper contributes to the big data research by categorizing the range of existing tools to measure the suggested dimensions, and to Library and Information Science (LIS) by proposing to account for heterogeneity of diverse big data, and to identify information quality dimensions important for each big data type.

show abstract

“…If the task is to classify the review as positive or negative, a two-step approach seems wise. The first step would be to filter out the objective sentences [Wiebe 2000;Wiebe et al 2001] and the second step would be to determine the semantic orientation of the words and phrases in the remaining subjective sentences [Turney 2002]. …”

Section: Subjectivity Analysismentioning

confidence: 99%

Measuring praise and criticism

Turney

Littman

2003

ACM Trans. Inf. Syst.

1,113

View full text Add to dashboard Cite

________________________________________________________________________The evaluative character of a word is called its semantic orientation. Positive semantic orientation indicates praise (e.g., "honest", "intrepid") and negative semantic orientation indicates criticism (e.g., "disturbing", "superfluous"). Semantic orientation varies in both direction (positive or negative) and degree (mild to strong). An automated system for measuring semantic orientation would have application in text classification, text filtering, tracking opinions in online discussions, analysis of survey responses, and automated chat systems (chatbots). This paper introduces a method for inferring the semantic orientation of a word from its statistical association with a set of positive and negative paradigm words. Two instances of this approach are evaluated, based on two different statistical measures of word association: pointwise mutual information (PMI) and latent semantic analysis (LSA). The method is experimentally tested with 3,596 words (including adjectives, adverbs, nouns, and verbs) that have been manually labeled positive (1,614 words) and negative (1,982 words). The method attains an accuracy of 82.8% on the full test set, but the accuracy rises above 95% when the algorithm is allowed to abstain from classifying mild words.

show abstract

A corpus study of evaluative and speculative language

Cited by 31 publications

References 22 publications

Deception Detection and Rumor Debunking for Social Media

Deception Detection and Rumor Debunking for Social Media

Veracity Roadmap: Is Big Data Objective, Truthful and Credible?

Measuring praise and criticism

Contact Info

Product

Resources

About