Social Media Comments Research Articles

Overlapping word removal is all you need: revisiting data imbalance in hope speech detection

ABSTRACT Hope speech detection is a new task for finding and highlighting positive comments or supporting content from user-generated social media comments. For this task, we have used a Shared Task multilingual dataset on Hope Speech Detection for Equality, Diversity, and Inclusion (HopeEDI) for three languages English, code-switched Tamil and Malayalam. In this paper, we present deep learning techniques using context-aware string embeddings for word representations and Recurrent Neural Network (RNN) and pooled document embeddings for text representation. We have evaluated and compared the three models for each language with different approaches. Our proposed methodology works fine and achieved higher performance than baselines. The highest weighted average F-scores of 0.93, 0.58, and 0.84 are obtained on the task organisers{'} final evaluation test set. The proposed models are outperforming the baselines by 3{\\%}, 2{\\%} and 11{\\%} in absolute terms for English, Tamil and Malayalam respectively.

Experiences of giving birth during the COVID-19 pandemic: a qualitative analysis of social media comments through the lens of birth integrity

BackgroundSocial media offer women a space to discuss birth-related fears and experiences. This is particularly the case during the COVID-19 pandemic when measures to contain the spread of the virus and high rates of infection have had an impact on the delivery of care, potentially restricting women’s rights and increasing the risk of experiencing different forms of mistreatment or violence. Through the lens of birth integrity, we focused on the experiences of women giving birth in Germany as shared on social media, and on what may have sheltered or violated their integrity during birth.MethodsUsing thematic analysis, we identified key themes in 127 comments and associated reactions (i.e. “likes”, emojis) posted on a Facebook public page in response to the dissemination of a research survey on maternity care in the first year of the COVID-19 pandemic.ResultsWomen contributing to the dataset gave birth during March and December 2020. They were most negatively affected by own mask-wearing –especially during the active phase of labour, not being allowed a birth companion of choice, lack of supportive care, and exclusion of their partner from the hospital. Those topics generated the most reactions, revealing compassion from other women and mixed feelings about health measures, from acceptation to anger. Many women explicitly formulated how inhumane or disrespectful the care was. While some women felt restricted by the tight visiting rules, those were seen as positive by others, who benefited from the relative quiet of maternity wards and opportunities for postpartum healing and bonding.ConclusionExceptional pandemic circumstances have introduced new parameters in maternity care, some of which appear acceptable, necessary, or beneficial to women, and some of which can be considered violations of birth integrity. Our research calls for the investigation of the long-term impact of those violations and the reassessment of the optimal conditions of the delivery of respectful maternity during the pandemic and beyond.

Multi-class sentiment classification on Bengali social media comments using machine learning

Multi-class Sentiment Analysis (SA) is an important field of computational linguistics that extracts multiple opinions expressed in a text using NLP and text-mining techniques. Existing research on multi-class SA in the Bengali language is directed towards ternary classification with unsatisfactory classification performance. Moreover, obtaining a higher performance score is challenging due to the peculiarities of Bengali text, lack of ground truth datasets, and low resources of preprocessing tools. Moreover, no research has shown that deep learning algorithms perform higher on four types of sentiments. Therefore, we proposed a supervised deep learning classifier based on CNN and LSTM to conduct multi-class SA on Bengali social media comments labelled as sexual, religious, political, and acceptable. The study aims to achieve maximum accuracy using the proposed model and provide a comparative analysis with the baseline models. Six machine learning models with two different feature extraction techniques were considered baseline models. The performance of our proposed CLSTM architecture can greatly improve the performance of SA with 85.8% accuracy and 0.86 F1 scores on a labelled dataset of 42,036 Facebook comments. A web application based on the proposed model and the highest-performing baseline model was built to detect the real-life sentiment of social media comments.

The unsanitary other and racism during the pandemic: analysis of purity discourses on social media in India, France and United States of America during the COVID-19 pandemic

The global rise of populism and concomitant polarizations across disenfranchised and marginalized groups has been magnified by so-called echo chambers, and a major public health crisis like the COVID-19 pandemic has only served to fuel these intergroup tensions. Media institutions disseminating information on ways to prevent the propagation of the virus have reactivated a specific discursive phenomenon previously observed in many epidemics: the construction of a defiled ‘Other’. With anthropological lenses, discourse on defilement is an interesting path to understand the continuous emergence of pseudo-scientific forms of racism. In this paper, the authors focus on ‘borderline racism’, that is the use of an institutionally ‘impartial’ discourse to reaffirm the inferiority of another race. The authors employed inductive thematic analysis of 1200 social media comments reacting to articles and videos published by six media in three different countries (France, United States and India). Results delineate four major themes structuring defilement discourses: food (and the relationship to animals), religion, nationalism and gender. Media articles and videos portrayed Western and Eastern countries through contrasting images and elicited a range of reaction in readers and viewers. The discussion reflects on how borderline racism can be an appropriate concept to understand the appearance of hygienic othering of specific subgroups on social media. Theoretical implications and recommendations on a more culturally sensitive approach of media coverage of epidemics and pandemics are discussed.

Comparative study on sentimental analysis using machine learning techniques

With the advancement of the Internet and the world wide web (WWW), it is observed that there is an exponential growth of data and information across the internet. In addition, there is a huge growth in digital or textual data generation. This is because users post the reply comments in social media websites based on the experiences about an event or product. Furthermore, people are interested to know whether the majority of potential buyers will have a positive or negative experience on the event or the product. This kind of classification in general can be attained through Sentiment Analysis which inputs unstructured text comments about the product reviews, events, etc., from all the reviews or comments posted by users. This further classifies the data into different categories namely positive, negative or neutral opinions. Sentiment analysis can be performed by different machine learning models like CNN, Naive Bayes, Decision Tree, XgBoost, Logistic Regression etc. The proposed work is compared with the existing solutions in terms of different performance metrics and XgBoost outperforms out of all other methods.

Identifying Fake Reviews in Relation with Property and Political Data Using Deep Learning

Modern deep learning algorithms have achieved tremendous success in many visual applications by training a model with all relevant task-specific data. In this module fake reviews evolving in online political and property data will be discussed. Regarding political information there are many differences while switching from offline to online mode. The views and comments of the politicians in social media act as a question mark “whether it is true or not?”. In some cases, it appears to be real and sometimes it may be a false/fake statement. All because, the online information is often considered as the most important source of info for the users these kinds of reviews and comments affects the trust factors of social media. Now days in social media, more of toxic contents and unwanted information were present rather than the useful information. These toxic comments acts as a threat for users using the application. Hence users do not come forward to post their thoughts and actions or to share the information. This module is used to distinguish reviews, identify the factor in online mode of political and property comments in social media. The LSTM and BERT (Bidirectional Encoder Representations from Transformers) algorithms are used in the first module because they offer a wide variety of parameters such as learning rates, input and output biases, and text categorization. Additionally, GPT2 (Generative Pre-Trained Transformer 2) is implemented which helps in text generation, increasing the size of dataset for training in different classification models. Thus it is important to understand for us that ultimately the model will be getting trained in such a way that it is able to give accurate results while we are undergoing classification of text and also while we are undergoing the identification of fake/authentic comments that are there in the source of our data.

The Role Of Social Media Comments In Clarifying The Meaning Of Posts And Its Impact On The Readability Of Automatic Translation

The Role Of Social Media Comments In Clarifying The Meaning Of Posts And Its Impact On The Readability Of Automatic Translation

Offensive language identification in dravidian languages using MPNet and CNN

Social media has effectively replaced traditional forms of communication and marketing. As these platforms allow for the free expression of ideas and facts through text, images, and videos, there exists a significant need to screen them to safeguard people and organisations from objectionable information directed at them. Our work aims to categorise code-mixed social media comments and posts in Tamil, Malayalam, and Kannada into offensive or not offensive at different levels. We present a multilingual MPNet and CNN fusion model for detecting offensive language content directed at an individual (or group) in low-resource Dravidian languages at different levels. Our model is capable of handling data that has been code-mixed, such as Tamil and Latin scripts. The model was successfully validated on the datasets, achieving offensive language detection results better than those of other baseline models with weighted average F1-score of 0.85, 0.98, and 0.76, and performed better than the baseline models EWDT, and EWODT by 0.02, 0.02, 0.04 for Tamil, Malayalam, and Kannada respectively.

A sentiment analysis approach to the prediction of market volatility.

Prediction and quantification of future volatility and returns play an important role in financial modeling, both in portfolio optimisation and risk management. Natural language processing today allows one to process news and social media comments to detect signals of investors' confidence. We have explored the relationship between sentiment extracted from financial news and tweets and FTSE100 movements. We investigated the strength of the correlation between sentiment measures on a given day and market volatility and returns observed the next day. We found that there is evidence of correlation between sentiment and stock market movements. Moreover, the sentiment captured from news headlines could be used as a signal to predict market returns; we also found that the same does not apply for volatility. However, for the sentiment found in Twitter comments we obtained, in a surprising finding, a correlation coefficient of -0.7 (p < 0.05), which indicates a strong negative correlation between negative sentiment captured from the tweets on a given day and the volatility observed the next day. It is important to keep in mind that stock volatility rises greatly when the market collapses but not symmetrically so when it goes up (the so-called leverage effect). We developed an accurate classifier for the prediction of market volatility in response to the arrival of new information by deploying topic modeling, based on Latent Dirichlet Allocation, in order to extract feature vectors from a collection of tweets and financial news. The obtained features were used as additional input to the classifier. Thanks to the combination of sentiment and topic modeling even on modest (essentially personal) architecture our classifier achieved a directional prediction accuracy for volatility of 63%.

A thematic analysis of YouTube comments on a television documentary titled 'Diabulimia: The World's most dangerous eating disorder'.

Omitting insulin for fear of weight gain is a type of disordered eating (also labelled as diabulimia) common in type 1 diabetes (T1D) and is associated with a worse biomedical prognosis but is not a formally recognised condition. This research explored the public's opinion of diabulimia as a condition as presented in a television documentary. We conducted a coding reliability thematic analysis using NVivo software of the original comments to a YouTube documentary 'Diabulimia: The World's Most Dangerous Eating Disorder' between 24 September 2017 and 16 June 2020. Of 1424 original comments, 1264 were eligible and uploaded into NVivo 12. The commenters were people with T1D, family and friends, health care professionals, and the wider public who collectively had questions, personal stories and/or opinions. Three main themes were discerned: lack of awareness of diabulimia as a condition; the importance of support; diabulimia as a psychiatric condition in the diabetes setting. This analysis of social media comments found that there is a lack of awareness of diabulimia amongst patients, their families and friends, and healthcare professionals and that there were many commenters who had the experience of disordered eating with T1D. This study has reported on themes that suggest there may be an eating disorder specifically in people with T1D and that further work is needed to understand the diagnostic criteria for diabulimia in order to develop effective treatments.

Unfair, Innocent, Flamed: Examining How the Chinese Public Perceived Sun Yang’s 8-Year Doping Sanction

International sport has always been associated with nationalism. The purpose of the study was to explore how Chinese media and the general public perceived the doping scandal of their national sports hero, Sun Yang. Through analyzing 11 Chinese media outlets’ coverage on Chinese social media Weibo, the results revealed that Chinese media covered Sun and his team’s reaction and perspectives on this issue more when compared with other news. The general public’s perceptions toward this scandal tended to be favorable toward Sun, with 55.5% of selected Weibo comments defending Sun after his 8-year ban for doping was handed down. The analysis of these social media comments posted by sports fans showed that the general public’s perceptions might have been impacted by their nationalism, international relations, and media coverage. In addition, the study revealed the Chinese public’s perceptions toward current antidoping regulations.

Digital Narratives of Living With Lupus: Lived Experiences and Meanings for Latin American and Latino Patients and Their Families.

Systemic lupus erythematosus (SLE) disproportionately affects Latin American and Latino populations, with worse outcomes compared to nonminority populations. Understanding patients' views is critical to provide culturally competent care. The objective of this research is to analyze lived experiences with SLE from comments made by Latin American and Latino patients, and their relatives and friends, on the public Facebook group "Hablemos de Lupus" (in English: "Let's Talk about Lupus"). Deidentified narratives posted as a reaction to the most popular resources shared by the page were extracted using the Facepager application. We conducted a thematic analysis under an interpretative medical anthropology framework. Five core themes were demonstrated by social media comments: lived experiences with lupus, religious/spiritual thoughts, metaphors, heredity, and experiences of family and friends. Being diagnosed with lupus is perceived as a life-changing event. The fluctuating course of the disease causes uncertainty, and the perception of invisibility within the patient's social circle generates feelings of being misunderstood. Faith and spiritual thoughts are coping strategies. Patients use metaphors about the disease's meaning and their lived experiences (the purple butterfly, not belonging, bellicose metaphors) to communicate with others. Relatives and friends are impacted by their loved one's distress. Patients perceive lupus as an unpredictable illness and use metaphors to foster empathy and communicate their experiences to others. Religion is as important as medical treatment to cope with the disease, and the experience of having lupus extends to family and friends. Findings can be used to improve physician-patient communication and lupus education campaigns in the Latin American and Latino population.

#waistgangsociety: Black Women’s Health Information—seeking Behaviors Concerning Body Modification Products Advertised on Instagram

ABSTRACT The purpose of this project is to determine the health information-seeking behaviors of Black women who use self-modification devices advertised on Instagram, and how they identify and negotiate risk factors related to these products. Forty-two interviews with Black women reveal that in seeking health risk information related to the devices, the women initially explored social media comments about the devices, replied with questions to the commenters, sent direct messages to people who indicated they bought the devices, and searched hashtags. They further clarified information beyond social media by searching Google, browsing websites like WebMD and Healthline, and reviewing accessible peer-reviewed sources. The health risk information was further confirmed with word-of-mouth firsthand experience or by testing the product on their own. Participants developed their own modification barometers to gauge what health impacts they were willing to accept in order to get the desired physical outcome. Advertisers or promoters can provide explicit and accurate details about the health risks of products, to provide transparency and valuable health information to consumers as they navigate their own modification barometers.

Empathy Gap in Social Media Comments for Sexual Harassment Victim

Indonesian Twitter users who filed a complaint about sexual harassment were studied to see if there was an empathy gap among their tweet responses. The author uses a content analysis method and observes a sample of empathy gap experiences to notice and study empathy gap behaviour in Twitter toward sexual harassment victims. In the research that has been done, the comments tweet as amount 3733 tweets and chosen 60 of them randomly to know-how is the empathy gap with sexual harassment cases. It is concluded that bullies have aggressive and intimidating characteristics. On message production by the bully, actors are supposed to produce messages in expressive, conventional, and rhetorical ways, including negative empathy characteristics. So on the other hand, the research that has been done concludes that people who act as victims have passive and defensive elements. On message reception by the communicant (victim), the victim placed the position of receiving the message in a dominant, negotiating, and oppositional position.

Evaluation of social media comments and scores of airline companies before and after Covid-19 with data and text mining

Veri ve metin madenciliği, anlamlı ilişkileri ve eğilimleri ayırt etmek için kullanıcıların taleplerine göre yapılandırılmış, yarı yapılandırılmış ve yapılandırılmamış büyük bir veri miktarını analiz etme sürecidir. İşletmeler, veri ve metin madenciliği teknikleri kullanarak hem kendi işletmeleri içerisinde hem de rakipleri ile rekabette yaşadıkları sorunlarına etkili çözümler üretebilmektedirler. Böylece elde ettikleri bilgiyi rekabet avantajına çevirebilmektedirler. Bu araştırmada, veri ve metin madenciliği algoritmaları kullanılarak rekabete dayalı pazarda müşterilerin istek ve ihtiyaçlarına göre hava yolu firmalarının üstün ve zayıf yönlerinin değerlendirilmesi amaçlanmıştır. Bu araştırmada, sosyal medya sitelerinden olan TripAdvisor’daki çevrimiçi seyahat incelemeleri araştırma kapsamına alınmıştır. Star Alliance küresel hava yolu birliğine üye 26 hava yolu firması değerlendirilmiştir. Araştırmada kullanılan kriterler; her bir kullanıcının yorum ve skorları temel alınarak belirlenmiştir. Duygu Analizi ile müşteri yorumlarından polariteleri belirlendikten sonra Destek Vektör Makineleri, Naive Bayes, Derin Öğrenme Algoritmaları ile sınıflandırma ve tahminleme yapılarak elde edilen sonuçlar karşılaştırılmıştır. Bunun yanı sıra sonuçlar, Covid-19 pandemisi öncesi ve sonrası olarak da karşılaştırılmıştır. Yapılan karşılaştırmada Derin Öğrenmenin daha iyi sonuç verdiği saptanmıştır.

The Evaluation of Rural Outdoor Dining Environment from Consumer Perspective.

The quality of the environment should be measured by the satisfaction of the public and guided by the issues of public concern. With the development of the internet, social media as the main platform for people to exchange information has become a data source for planning and management analysis. Nowadays, the rural catering industry is becoming increasingly competitive, especially after the pandemic. How to further enhance the competitiveness of the rural catering industry has become a hot topic in the industry. From the perspective of consumers, we explored consumers' preferences in a rural outdoor dining environment through social media data. The research analyzed the social media data through manual collection and object detection, divided the landscape of the rural outdoor dining environment into eight categories with 35 landscape elements, and then used BP (Back Propagation) neural network nonlinear fitting and least square linear fitting to analyze the 11,410 effective review pictures from eight rural restaurants' social media comments in Chengdu. We derived the degree of consumer preference for the landscape quality of the rural outdoor dining environment and analyzed the differences in preference among three different groups (regular customers, customers with children, and customers with the elderly). The study found that agricultural resources are an important factor in the competitiveness of rural restaurant environments; that children's emotions when using activity facilities can positively influence consumers' dining experiences; that safety and hygiene environment are important factors influencing the decisions of parent-child dining; and that older people are more interested in outdoor nature, etc. The research results provide suggestions and knowledge for rural restaurant managers and designers through human-oriented needs from the perspective of consumers, and clarify the preferences and expectations of different consumer groups for rural restaurant landscapes while achieving the goal of rural landscape protection.

From ‘Main Course’ to ‘Side Dish?’ Debates about Removing English as a Core Subject for Chinese Students Receiving Compulsory Education

ABSTRACT China’s impressive growth over the last three decades warrants the need to re-evaluate the position and positioning of English in the country’s educational system. For long, English has been taught and learned compulsorily alongside Chinese and maths in primary and secondary schools in the mainland of China. It was not until recent years that the legitimacy of English as a core subject started to be questioned, and calls for alternative solutions began to emerge. For a better understanding of the changing role of English in China’s compulsory education and its influence on stakeholders and policymakers, this article investigates the public reaction to a recent controversial proposal to remove English as a core subject from the national curriculum. Drawing from social media comments, autoethnographic reflections, and existing research, the study presents major opinions and attitudes that feature complex sentiments and ambivalences intersected by concerns about equity and balance.

Attitudes towards the 'Shisha No Thanks' campaign video: Content analysis of Facebook comments.

While social media are commonly used in public health campaigns, there is a gap in our understanding of what happens after the campaign is seen by the target audience. This study aims to understand how the Shisha No Thanks campaign video was received by the Facebook audience by analyzing Facebook comments posted to it. Specifically, this study aims to determine whether the Facebook audience accepted or rejected the campaign's message. A sample of the Facebook comments was extracted, and the study team, which included cultural support workers, developed content categories consistent with the research question. Each comment was then coded by three team members, and only assigned a category if there was agreement by at least two members. Of the 4990 comments that were sampled, 9.1% (456) accepted the campaign message, 22.9% (1144) rejected the message, 21.8% (1089) were unclear, and 46.1% (2301) contained only tagged names. Of the sample, 2.8% (138) indicated the commenter took on board the campaign message by expressing an intention to stop smoking shisha, or asking a friend to stop smoking shisha. Of the comments that showed rejection of the campaign, the majority were people dismissing the campaign by laughing at it or expressing pro-shisha sentiments. This study demonstrates that conducting content analyses of social media comments can provide important insight into how a campaign message is received by a social media audience.

Consumers' Opinions towards Public Health Effects of Online Games: An Empirical Study Based on Social Media Comments in China.

Online game products have fueled the boom in China’s digital economy. Meanwhile, its public health concerns have sparked discussion among consumers on social media. However, past research has seldom studied the public health topics caused by online games from the perspective of consumer opinions. This paper attempts to identify consumers’ opinions on the health impact of online game products through non-structured text and large-size social media comments. Thus, we designed a natural language processing (NLP) framework based on machine learning, which consists of topic mining, multi-label classification, and sentimental analysis. The hierarchical clustering method-based topic mining procedure determines the compatibility of this study and previous research. Every three topics are identified in “Personal Health Effects” and “Social Health Effects”, respectively. Then, the multi-label classification model’s results show that 61.62% of 327,505 comments have opinions about the health effects of online games. Topics “Adolescent Education” and “Commercial Morality” occupy the top two places of consumer attention. More than 31% of comments support two or more topics, and the “Adolescent Education” and “Commercial Morality” combination also have the highest co-occurrence. Finally, consumers expressed different emotional preferences for different topics, with an average of 63% of comments expressing negative emotions related to the health attributes of online games. In general, Chinese consumers are most concerned with adolescent education issues and hold the strongest negative emotion towards the commercial morality problems of enterprises. The significance of research results is that it reminds online game-related enterprises to pay attention to the potential harm to public health while bringing about additional profits through online game products. Furthermore, negative consumer emotions may cause damage to brand image, business reputation, and the sustainable development of the enterprises themselves. It also provides the government supervision departments with an advanced analysis method reference for more effective administration to protect public health and promote the development of the digital economy.

Time for different stories: Reflections on IPONS panel addressing current debates in nursing theory, educationand practice.

On 4 February 2021 a group of nurse scholar-educators, nurses and other interested folks came together for the second of two virtual events to think together about the role of philosophy in the nursing world. The live streamed open access event provided an opportunity in the COVID-19 pandemic for over 400 people to listen to five nursing scholars' presentations and to interact virtually through comments in chat and on the @IPONSociety Twitter social media platform. By reading the comments and questions that were generated, and by looking at the social media comments related to the event, it is apparent that philosophy is an important thinking practice for nurses but many audience members critically expressed they felt excluded. Critical issues were raised by participants in chat and on Twitter-pointedly around the need for more representative voices-including the imperative to open nursing philosophy to diverse and disparate worldviews. This dialogue provides a summary of critical points raised during the live question and answer session for the panel entitled Addressing Current Debates in Nursing Theory, Education, Practice as well as examing comments selected from the @IPONSociety Twitter space in response to the panel. One commenter said it was great to see the discussion being lifted up from the influential roots of white supremacy, while other nurses expressed that they wished the panellists themselves were more diverse. In discussion of key takeaway, links are made to historical and ongoing structural oppressions in nursing where thinking practices like nursing philosophy and theory are still dominated by world views emanating from positionalities of able-bodied cis-gendered heterosexual western eurocentric whiteness.