Detecting Sarcasm is Extremely Easy ;-)

Parde, Natalie; Nielsen, Rodney D.

doi:10.18653/v1/w18-1303

Cited by 15 publications

(9 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Social media messages, especially from strictly character restricted platforms such as Twitter, are of primary concern, given their often short sparse nature and lack of context. This is to some extent empirically illustrated in the work by Parde and Nielsen (2018), where sarcasm detection over Amazon reviews, compared to Twitter, is substantially more accurate due to the longer length of content and hence richer context available to the sarcasm detection model. Another implication stems from our observation of a very high proportion of url links (80%) seen among humorous tweets, as well as sarcastic (41.56%) and ironic tweets (49.50%).…”

Section: Discussionmentioning

confidence: 88%

“…used sarcasm hashtags to generate their seed dataset, which is a method that, as we show in 'Prevalence of sarcasm and irony: A manual semantic analysis' section in this paper, is rather problematic. Parde and Nielsen (2018) have shown that using sarcastic tweets (the #sarcasm hashtag) and enriching training datasets with additional annotated Amazon product review data can slightly improve the performance of automated sarcasm detection, where the F score improvements reported are 0.58 to 0.59 for tweets and 0.74 to 0.78 for Amazon reviews. Amazon-based sarcasm detection is more accurate than detection on Twitter, primarily due to longer messages.…”

Section: Sarcasmmentioning

confidence: 99%

“…Using sarcasm and irony hashtags as "gold" 8 labels for training sarcasm detection systems is a widely used method (e.g., Bamman and Smith, 2015;Khattri et al, 2015;Parde and Nielsen, 2018;Rajadesingan et al, 2015;Sarsam et al, 2020). Although distant supervision-based learning was found to work well for basic sentiment classification (Hogenboom et al, 2013;Pak and Paroubek, 2010), doubts about such approaches being used for sarcasm and irony detection have been raised some time ago by Davidov et al (2010).…”

Section: Prevalence Of Sarcasm and Irony: A Manual Semantic Analysismentioning

confidence: 99%

“…Hee et al (2018) have specifically shown how introducing knowledge bases of common-sense, also known as general world-knowledge semantics, can significantly improve performance in detecting irony (in their case achieving a respectable F-measure of just over 0.7). Indeed, an error analysis of misclassifications for automated sarcasm detection by Parde and Nielsen (2018) highlights three particular issues and areas for future work; (i) introducing models of world or common-sense knowledge to help detect sarcasm/ irony, (ii) better approaches to normalising text, such as splitting up compound hashtags into individual words, to ultimately improving performance of sentiment detection approaches by appropriately reversing the polarity of detected sentiment, and (iii) issues around evolving language and slang on social media pose a continued challenge. These three future research challenge areas directly relate to also addressing the multi-word hashtag issue we have pointed out above.…”

Section: Future Workmentioning

confidence: 99%

See 3 more Smart Citations

A qualitative analysis of sarcasm, irony and related #hashtags on Twitter

Sýkora¹,

Elayan²,

Jackson

2020

Big Data & Society

View full text Add to dashboard Cite

As the use of automated social media analysis tools surges, concerns over accuracy of analytics have increased. Some tentative evidence suggests that sarcasm alone could account for as much as a 50% drop in accuracy when automatically detecting sentiment. This paper assesses and outlines the prevalence of sarcastic and ironic language within social media posts. Several past studies proposed models for automatic sarcasm and irony detection for sentiment analysis; however, these approaches result in models trained on training data of highly questionable quality, with little qualitative appreciation of the underlying data. To understand the issues and scale of the problem, we are the first to conduct and present results of a focused manual semantic annotation analysis of two datasets of Twitter messages (in total 4334 tweets), associated with; (i) hashtags commonly employed in automated sarcasm and irony detection approaches, and (ii) tweets relating to 25 distinct events, including, scandals, product releases, cultural events, accidents, terror incidents, etc. We also highlight the contextualised use of multi-word hashtags in the communication of humour, sarcasm and irony, pointing out that many sentiment analysis tools simply fail to recognise such hashtag-based expressions. Our findings also offer indicative evidence regarding the quality of training data used for automated machine learning models in sarcasm, irony and sentiment detection. Worryingly only 15% of tweets labelled as sarcastic were truly sarcastic. We highlight the need for future research studies to rethink their approach to data preparation and a more careful interpretation of sentiment analysis.

show abstract

Section: Discussionmentioning

confidence: 88%

Section: Sarcasmmentioning

confidence: 99%

Section: Prevalence Of Sarcasm and Irony: A Manual Semantic Analysismentioning

confidence: 99%

Section: Future Workmentioning

confidence: 99%

See 2 more Smart Citations

A qualitative analysis of sarcasm, irony and related #hashtags on Twitter

Sýkora¹,

Elayan²,

Jackson

2020

Big Data & Society

View full text Add to dashboard Cite

show abstract

“…Many previous studies relied on the application of Naïve Bayes for various detection purposes. In the context of this study, Naïve Bayes was commonly used to classify tweets’ polarity (positive and negative; Parde & Nielsen, 2018). Another way of labeling was established by Das et al (2018) who utilized two types of labels (0 and 1) by indicating whether the tweet is sarcastic or not sarcastic.…”

Section: Amlamentioning

confidence: 99%

Sarcasm detection using machine learning algorithms in Twitter: A systematic review

Sarsam

Al-Samarraie

Alzahrani

et al. 2020

International Journal of Market Research

View full text Add to dashboard Cite

Recognizing both literal and figurative meanings is crucial to understanding users’ opinions on various topics or events in social media. Detecting the sarcastic posts on social media has received much attention recently, particularly because sarcastic comments in the form of tweets often include positive words that represent negative or undesirable characteristics. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement was used to understand the application of different machine learning algorithms for sarcasm detection in Twitter. Extensive database searching led to the inclusion of 31 studies classified into two groups: Adapted Machine Learning Algorithms (AMLA) and Customized Machine Learning Algorithms (CMLA). The review results revealed that Support Vector Machine (SVM) was the best and the most commonly used AMLA for sarcasm detection in Twitter. In addition, combining Convolutional Neural Network (CNN) and SVM was found to offer a high prediction accuracy. Moreover, our result showed that using lexical, pragmatic, frequency, and part-of-speech tagging can contribute to the performance of SVM, whereas both lexical and personal features can enhance the performance of CNN-SVM. This work also addressed the main challenges faced by prior scholars when predicting sarcastic tweets. Such knowledge can be useful for future researchers or machine learning developers to consider the major issues of classifying sarcastic posts in social media.

show abstract