Search results
1 – 10 of 762Daniel Šandor and Marina Bagić Babac
Sarcasm is a linguistic expression that usually carries the opposite meaning of what is being said by words, thus making it difficult for machines to discover the actual meaning…
Abstract
Purpose
Sarcasm is a linguistic expression that usually carries the opposite meaning of what is being said by words, thus making it difficult for machines to discover the actual meaning. It is mainly distinguished by the inflection with which it is spoken, with an undercurrent of irony, and is largely dependent on context, which makes it a difficult task for computational analysis. Moreover, sarcasm expresses negative sentiments using positive words, allowing it to easily confuse sentiment analysis models. This paper aims to demonstrate the task of sarcasm detection using the approach of machine and deep learning.
Design/methodology/approach
For the purpose of sarcasm detection, machine and deep learning models were used on a data set consisting of 1.3 million social media comments, including both sarcastic and non-sarcastic comments. The data set was pre-processed using natural language processing methods, and additional features were extracted and analysed. Several machine learning models, including logistic regression, ridge regression, linear support vector and support vector machines, along with two deep learning models based on bidirectional long short-term memory and one bidirectional encoder representations from transformers (BERT)-based model, were implemented, evaluated and compared.
Findings
The performance of machine and deep learning models was compared in the task of sarcasm detection, and possible ways of improvement were discussed. Deep learning models showed more promise, performance-wise, for this type of task. Specifically, a state-of-the-art model in natural language processing, namely, BERT-based model, outperformed other machine and deep learning models.
Originality/value
This study compared the performance of the various machine and deep learning models in the task of sarcasm detection using the data set of 1.3 million comments from social media.
Details
Keywords
Phoey Lee Teh, Pei Boon Ooi, Nee Nee Chan and Yee Kang Chuah
Sarcasm is often used in everyday speech and writing and is prevalent in online contexts. The purpose of this paper is to investigate the analogy between sarcasm comments from…
Abstract
Purpose
Sarcasm is often used in everyday speech and writing and is prevalent in online contexts. The purpose of this paper is to investigate the analogy between sarcasm comments from sentiment tools and the human coder.
Design/methodology/approach
Using the Verbal Irony Procedure, eight human coders were engaged to analyse comments collected from an online commercial page, and a dissimilarity analysis was conducted with sentiment tools. Three constants were tested, namely, polarity from sentiment tools, polarity rating by human coders; and sarcasm-level ratings by human coders.
Findings
Results found an inconsistent ratio between these three constants. Sentiment tools used did not have the capability or reliability to detect the subtle, contextualized meanings of sarcasm statements that human coders could detect. Further research is required to refine the sentiment tools to enhance their sensitivity and capability.
Practical implications
With these findings, it is recommended that further research and commercialization efforts be directed at improving current sentiment tools – for example, to incorporate sophisticated human sarcasm texts in their analytical systems. Sarcasm exists frequently in media, politics and human forms of communications in society. Therefore, more highly sophisticated sentiment tools with the abilities to detect human sarcasm would be vital in research and industry.
Social implications
The findings suggest that presently, of the sentiment tools investigated, most are still unable to pick up subtle contexts within the text which can reverse or change the message that the writer intends to send to his/her receiver. Hence, the use of the relevant hashtags (e.g. #sarcasm; #irony) are of fundamental importance in detection tools. This would aid the evaluation of product reviews online for commercial usage.
Originality/value
The value of this study lies in its original, empirical findings on the inconsistencies between sentiment tools and human coders in sarcasm detection. The current study proves these inconsistencies are detected between human and sentiment tools in social media texts and points to the inadequacies of current sentiment tools. With these findings, it is recommended that further research and commercialization efforts be directed at improving current sentiment tools – to incorporate sophisticated human sarcasm texts in their analytical systems. The system can then be used as a reference for psychologists, media analysts, researchers and speech writers to detect cues in the inconsistencies in behaviour and language.
Details
Keywords
Shubhadeep Mukherjee and Pradip Kumar Bala
The purpose of this paper is to study sarcasm in online text – specifically on twitter – to better understand customer opinions about social issues, products, services, etc. This…
Abstract
Purpose
The purpose of this paper is to study sarcasm in online text – specifically on twitter – to better understand customer opinions about social issues, products, services, etc. This can be immensely helpful in reducing incorrect classification of consumer sentiment toward issues, products and services.
Design/methodology/approach
In this study, 5,000 tweets were downloaded and analyzed. Relevant features were extracted and supervised learning algorithms were applied to identify the best differentiating features between a sarcastic and non-sarcastic sentence.
Findings
The results using two different classification algorithms, namely, Naïve Bayes and maximum entropy show that function words and content words together are most effective in identifying sarcasm in tweets. The most differentiating features between a sarcastic and a non-sarcastic tweet were identified.
Practical implications
Understanding the use of sarcasm in tweets let companies do better sentiment analysis and product recommendations for users. This could help businesses attract new customers and retain the old ones resulting in better customer management.
Originality/value
This paper uses novel features to identify sarcasm in online text which is one of the most challenging problems in natural language processing. To the authors’ knowledge, this is the first study on sarcasm detection from a customer management perspective.
Details
Keywords
Pradeep Kumar and Gaurav Sarin
Sarcasm is a sentiment in which human beings convey messages with the opposite meanings to hurt someone emotionally or condemn something in a witty manner. The difference between…
Abstract
Purpose
Sarcasm is a sentiment in which human beings convey messages with the opposite meanings to hurt someone emotionally or condemn something in a witty manner. The difference between the text's literal and its intended meaning makes it tough to identify. Mostly, researchers and practitioners only consider explicit information for text classification; however, considering implicit with explicit information will enhance the classifier's accuracy. Several sarcasm detection studies focus on syntactic, lexical or pragmatic features that are uttered using words, emoticons and exclamation marks. Discrete models, which are utilized by many existing works, require manual features that are costly to uncover.
Design/methodology/approach
In this research, word embeddings used for feature extraction are combined with context-aware language models to provide automatic feature engineering capabilities as well superior classification performance as compared to baseline models. Performance of the proposed models has been shown on three benchmark datasets over different evaluation metrics namely misclassification rate, receiver operating characteristic (ROC) curve and area under curve (AUC).
Findings
Experimental results suggest that FastText word embedding technique with BERT language model gives higher accuracy and helps to identify the sarcastic textual element correctly.
Originality/value
Sarcasm detection is a sub-task of sentiment analysis. To help in appropriate data-driven decision-making, the sentiment of the text that gets reversed due to sarcasm needs to be detected properly. In online social environments, it is critical for businesses and individuals to detect the correct sentiment polarity. This will aid in the right selling and buying of products and/or services, leading to higher sales and better market share for businesses, and meeting the quality requirements of customers.
Details
Keywords
Jyoti Godara, Rajni Aron and Mohammad Shabaz
Sentiment analysis has observed a nascent interest over the past decade in the field of social media analytics. With major advances in the volume, rationality and veracity of…
Abstract
Purpose
Sentiment analysis has observed a nascent interest over the past decade in the field of social media analytics. With major advances in the volume, rationality and veracity of social networking data, the misunderstanding, uncertainty and inaccuracy within the data have multiplied. In the textual data, the location of sarcasm is a challenging task. It is a different way of expressing sentiments, in which people write or says something different than what they actually intended to. So, the researchers are showing interest to develop various techniques for the detection of sarcasm in the texts to boost the performance of sentiment analysis. This paper aims to overview the sentiment analysis, sarcasm and related work for sarcasm detection. Further, this paper provides training to health-care professionals to make the decision on the patient’s sentiments.
Design/methodology/approach
This paper has compared the performance of five different classifiers – support vector machine, naïve Bayes classifier, decision tree classifier, AdaBoost classifier and K-nearest neighbour on the Twitter data set.
Findings
This paper has observed that naïve Bayes has performed the best having the highest accuracy of 61.18%, and decision tree performed the worst with an accuracy of 54.27%. Accuracy of AdaBoost, K-nearest neighbour and support vector machine measured were 56.13%, 54.81% and 59.55%, respectively.
Originality/value
This research work is original.
Details
Keywords
Daniel Kusaila and Natalie Gerhart
Technology-enabled communication used in workplace settings includes nuanced tools such as emojis, that are interpreted differently by different populations of people. This paper…
Abstract
Purpose
Technology-enabled communication used in workplace settings includes nuanced tools such as emojis, that are interpreted differently by different populations of people. This paper aims to evaluate the use of emojis in work environments, particularly when they are used sarcastically.
Design/methodology/approach
This research uses a survey method administered on MTurk. Overall, 200 participants were included in the analysis. Items were contextualized from prior research and offered on a seven-point Likert scale.
Findings
Females are better able to understand if an emoji is used sarcastically. Additionally, older employees are more capable of interpreting sarcasm than younger employees. Finally, understanding of emojis has a negative relationship with frustration, indicating that when users understand emojis are being used sarcastically, frustration is reduced.
Research limitations/implications
This research is primarily limited by the survey methodology. Despite this, it provides implications for theory of mind and practical understanding of emoji use in professional settings. This research indicates emojis are often misinterpreted in professional settings.
Originality/value
The use of emojis is becoming commonplace. The authors show the use of emojis in a professional setting creates confusion, and in some instances can lead to frustration. This work can help businesses understand how best to manage employees with changing communication tools.
Details
Keywords
Yung-Cheng Shen, Crystal T. Lee and Wen-Ya Lin
The proliferation of digital communication on social media provides new opportunities for businesses to take advantage of Internet memes to boost customer engagement. Academic…
Abstract
Purpose
The proliferation of digital communication on social media provides new opportunities for businesses to take advantage of Internet memes to boost customer engagement. Academic literature on digital communications mostly focuses on popular forms such as selfies, branded posts, and branded emoticons. Less attention has been paid to brand memes and their implications for brand management. Based on the cue utilization theory, this research aims to investigate the informational cues of brand memes foster brand partnerships.
Design/methodology/approach
The structural equation modeling and importance-performance matrix analysis were used to empirically validate the research hypotheses with 595 respondents to an online survey.
Findings
Three informational cues of brand memes (i.e. comprehensibility, novelty, and meme-brand congruity) stimulated consumers' attitudes, which in turn impacted consumer-brand relationships. Another brand meme informational cue, sarcasm, negatively moderated the relationships between the three informational cues and consumer-brand relationships.
Originality/value
Our findings indicate that a brand can engage consumers in conversations on social media and foster long-term consumer-brand relationships through brand memes.
Details
Keywords
Purpose: Previous research identified a measurement gap in the individual assessment of social misconduct in the workplace related to gender. This gap implies that women respond…
Abstract
Purpose: Previous research identified a measurement gap in the individual assessment of social misconduct in the workplace related to gender. This gap implies that women respond to comparable self-reported acts of bullying or sexual discrimination slightly more often than men with the self-labeling as “bullied” or “sexually discriminated and/or harassed.” This study tests this hypothesis for women and men in the scientific workplace and explores patterns of gender-related differences in self-reporting behavior.
Basic design: The hypotheses on the connection between gender and the threshold for self-labeling as having been bullied or sexually discriminated against were tested based on a sample from a large German research organization. The sample includes 5,831 responses on bullying and 6,987 on sexual discrimination (coverage of 24.5 resp. 29.4 percentage of all employees). Due to a large number of cases and the associated high statistical power, this sample for the first time allows a detailed analysis of the “gender-related measurement gap.” The research questions formulated in this study were addressed using two hierarchical regression models to predict the mean values of persons who self-labeled as having been bullied or sexually discriminated against. The status of the respondents as scientific or non-scientific employees was included as a control variable.
Results: According to a self-labeling approach, women reported both bullying and sexual discrimination more frequently. This difference between women and men disappeared for sexual discrimination when, in addition to the gender of a person, self-reported behavioral items were considered in the prediction of self-labeling. For bullying, the difference between the two genders remained even in this extended prediction. No statistically significant relationship was found between the frequency of self-reported items and the effect size of their interaction with gender for either bullying or sexual discrimination. When comparing bullying and sexual discrimination, it should be emphasized that, on average, women report experiencing a larger number of different behavioral items than men.
Interpretation and relevance: The results of the study support the current state of research. However, they also show how volatile the measurement instruments for bullying and sexual discrimination are. For example, the gender-related measurement gap is considerably influenced by single items in the Negative Acts Questionnaire and Sexual Experience Questionnaire. The results suggest that women are generally more likely than men to report having experienced bullying and sexual discrimination. While an unexplained “gender gap” in the understanding of bullying was found for bullying, this was not the case for sexual discrimination.
Details
Keywords
This study attempts to identify and analyze the pragmatic functions of religious expressions, that is, invocations that include the name of Allah (God), in naturally occurring…
Abstract
Purpose
This study attempts to identify and analyze the pragmatic functions of religious expressions, that is, invocations that include the name of Allah (God), in naturally occurring social interactions in Najdi Arabic, which is spoken in Central Saudi Arabia.
Design/methodology/approach
Drawing on the speech act theory and politeness model, an analysis of the data illustrates that religious expressions, in addition to their prototypical religious meanings and uses in everyday interactions, are employed to communicate a wide range of pragmatic functions.
Findings
These include signaling the end of a conversation, persuading, mitigating and hedging, showing agreement and approval, reinforcing emphasis, expressing emotions, seeking protection from the evil eye, conveying skepticism and ambiguity, expressing humor and sarcasm, and showing respect and honor. The embedded multifunctional dimension of religious expressions in the present data is interpreted as serving as a politeness marker with which speakers promote both positive politeness (by showing solidarity, claiming common grounds, and building rapport) and negative politeness (by reducing imposition and emphasizing personal autonomy).
Originality/value
This study further highlights the interplay between religion, culture, and language use in Najdi Arabic.
Details
Keywords
Pedro Quelhas Brito, Sandra Torres and Jéssica Fernandes
The purpose of this paper is to study the nature and concept of emoticons/emojis. Instead of taking for granted that these user-generated formats are necessarily emotional, we…
Abstract
Purpose
The purpose of this paper is to study the nature and concept of emoticons/emojis. Instead of taking for granted that these user-generated formats are necessarily emotional, we empirically assessed in what extent are they and the specificity of each one. Drawing on congruent mood state, valence core and emotion appraisal theories we expected a compatible statistical association between positive/negative/neutral emotional valence expressions and emoticons of similar valence. The positive emoticons were consistently associated with positive valence posts. Added to that analysis, 21 emotional categories were identified in posts and correlated with eight emoticons.
Design/methodology/approach
Two studies were used to address this question. The first study defined emoticon concept and interpreted their meaning highlighting their communication goals and anticipated effects. The link between emojis and emoticons was also obtained. Some emoticons types present more ambiguity than others. In the second study, three years of real and private (Facebook) posts from 82 adolescents were content analyzed and coded.
Findings
Only the neutral emoticons always matched neutral emotional categories found in the written interaction. Although the emoticon valence and emotional category congruence pattern was the rule, we also detected a combination of different valence emoticons types and emotion categories valence expressions. Apparently the connection between emoticon and emotion are not so obviously straightforward as the literature used to assume. The created objects designed to communicate emotions (emoticons) have their specific corresponding logic with the emotional tone of the message.
Originality/value
Theoretically, we discussed the emotional content of emoticons/emojis. Although this king of signals have an Asian origin and later borrowed from the western countries, their ambiguity and differing specificity have never been analyzed.
Details