Finding Deceptive Opinion Spam by Any Stretch of the Imagination

By Myle Ott, Yejin Choi, Claire Cardie, Jeffrey T. Hancock. 2011

1,500 citations recorded by the source corpus.

Consumers increasingly rate, review and research products online (Jansen, 2010; Litvin et al., 2008). Consequently, websites containing consumer reviews are becoming targets of opinion spam. While recent work has focused primarily on manually identifiable instances of opinion spam, in this work we study deceptive opinion spam---fictitious opinions that have been deliberately written to sound authentic. Integrating work from psychology and computational linguistics, we develop and compare three approaches to detecting deceptive opinion spam, and ultimately develop a classifier that is nearly 90% accurate on our gold-standard opinion spam dataset. Based on feature analysis of our learned models, we additionally make several theoretical contributions, including revealing a relationship between deceptive opinions and imaginative writing.

Subfields: Artificial Intelligence, Information Systems, Sociology and Political Science

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Sociology and Political Science

Citations in the corpus

The Effect of Different Writing Tasks on Linguistic Style: A Case Study of the ROC Story Cloze Task (2017, background): “Several researchers have used stylometric features to predict deception (Newman et al., 2003; Hancock et al., 2007; Ott et al., 2011; Feng et al., 2012).”
The Social Impact of Natural Language Processing (2016, methodology): “…automatic inference of user attributes, a common and interesting NLP task, whose solution also holds promise for many useful applications, such as recommendation engines and fraud or deception detection (Badaskar et al., 2008; Fornaciari and Poesio, 2014; Ott et al., 2011; Banerjee et al., 2014).”
“Liar, Liar Pants on Fire”: A New Benchmark Dataset for Fake News Detection (2017, background): “A relatively early study by Ott et al. (2011) focuses on detecting deceptive review opinions in sentiment analysis, using a crowdsourcing approach to create training data for the positive class, and then combine with truthful opinions from TripAdvisor.”
Automated Fact Checking: Task Formulations, Methods and Future Directions (2018, methodology): “Ott et al. (2011) incorporate the use of psycholinguistic cues to improve classiﬁcation accuracy.”
Opinion Fraud Detection in Online Reviews by Network Effects (2013, methodology): “Previous attempts at solving the problem use several heuristics, such as duplicated reviews (Jindal and Liu 2008), or acquire bogus reviews from non-experts (Ott et al. 2011), to generate pseudo-ground truth, or a reference dataset.”
Modelling users' trust in online health rumours: an experiment-based study in China (2021, background): “Informativeness has been identified as a key dimension of information quality (Doll and Torkzadeh, 1988) and as having an impact on trusting beliefs (Banerjee and Chua, 2014; Newman et al., 2003; Ott et al., 2011; Vrij et al., 2000).”
Man vs. Machine: Practical Adversarial Detection of Malicious Crowdsourcing Workers (2014, background): “Sybils in social networks [52, 42], scammers in e-commerce sites [53] and fraudulent reviewers on online review sites [31].”
Spotting opinion spammers using behavioral footprints (2013, methodology): “It proposes a novel and principled method to exploit observed behavioral footprints to detect spammers (fake reviewers) in an unsupervised Bayesian framework precluding the need of any manual labels for learning which is both hard [14] and noisy [34].”
Review Graph Based Online Store Review Spammer Detection (2011, background): “Our second contribution is the introduction of three fundamental concepts, i.e. the trustiness of reviewers, the honesty of reviews, and the reliability of stores, and the identification of their interrelationships: a reviewer is more trustworthy if (s)he has written more honesty reviews, a store…”
False Information on Web and Social Media: A Survey (2018, methodology): “[70] collected 400 truthful and 400 positive-sentiment deceptive AMT-sourced reviews and trained Support Vector Machine (SVM) classifiers using a variety of feature sets, such as n-grams and LIWC features [72].”
Stylometric Analysis of Scientific Articles (2012, background): “Inadditiontocontrastingdifferentfeaturetypes,wealsocomparedifferent , 2011), or whether someone is writing deceptive online reviews (Ott et al., 2011).”
Sentiment Analysis: An Overview from Linguistics (2016, neutral): “General surveys from a computational point of view are presented by Pang and Lee (2008), Liu (2012) and Sonntag and Stede (2014), whereas Feldman (2013) is a short overview for a lay”
The manager's dilemma: a conceptualization of online review manipulation strategies (2018, background): “Based on an understanding that manipulation does exist, (Yoo&Gretzel, 2009; Ott et al., 2011), Banerjee and Chua (2014a) investigated whether linguistic cues can reveal if a review is authentic ormanipulated.”