Multiple Social Media Sites Research Articles

Deep Learning Based Cyberbullying Detection in Bangla Language

The Internet is currently the largest platform for global communication including expressions of opinions, reviews, contents, images, videos and so forth. Moreover, social media has now become a very broad and highly engaging platform due to its immense popularity and swift adoption trend. Increased social networking, however, also has detrimental impacts on the society leading to a range of unwanted phenomena, such as online assault, intimidation, digital bullying, criminality and trolling. Hence, cyberbullying has become a pervasive and worrying problem that poses considerable psychological and emotional harm to the people, particularly amongst the teens and the young adults. In order to lessen its negative effects and provide victims with prompt support, a great deal of research to identify cyberbullying instances at various online platforms is emerging. In comparison to other languages, Bangla (also known as Bengali) has fewer research studies in this domain. This study demonstrates a deep learning strategy for identifying cyberbullying in Bengali, using a dataset of 12282 versatile comments from multiple social media sites. In this study, a two-layer bidirectional long short-term memory (Bi-LSTM) model has been built to identify cyberbullying, using a variety of optimisers as well as 5-fold cross validation. To evaluate the functionality and efficacy of the proposed system, rigorous assessment and validation procedures have been employed throughout the project. The results of this study reveals that the proposed model’s accuracy, using momentum-based stochastic gradient descent (SGD) optimiser, is 94.46%. It also reflects a higher accuracy of 95.08% and a F1 score of 95.23% using Adam optimiser as well as a better accuracy of 94.31% in 5-fold cross validation.

CRC-PREVENT: Clinical validation trial to show expedited and diverse recruitment for the non-invasive RNA-FIT test that can detect advanced colorectal neoplasia with high sensitivity.

23 Background: Prevailing methods for patient recruitment in large prospective studies can be time consuming, expensive, and introduce selection bias against patients with low health literacy or reduced access to healthcare. Previous clinical trials have reported low recruitment of women, minorities, and individuals who face socioeconomic barriers; a concern which has been exacerbated by the COVID-19 pandemic. Here we describe a novel recruitment strategy that helps to address healthcare disparities. This study will support a pre-market approval application to the FDA for a multi-factor RNA-FIT assay for detection of colorectal neoplasia in average-risk individuals between the ages of 45-75. Methods: A decentralized clinical trial (CRC-PREVENT) was launched through a digital campaign (https://www.colonscreeningstudy.com/; NCT04739722) after the RNA-FIT test system entered design-lock. Online advertisements were published on multiple social media sites and engagement with materials directed patients to an online screener. Participants who completed the screener were considered eligible for enrollment if they met CRC-PREVENT inclusion/ exclusion criteria and were willing to complete all components of the clinical trial, including providing a stool sample prior to an optical colonoscopy. Results: After 3 months of active enrollment, 51,588 individuals have engaged with digital advertisements and completed pre-screener surveys to determine eligibility. In total, 35,280 individuals were deemed eligible based on survey response, and 13,294 eligible individuals also expressed interest in the CRC-PREVENT clinical trial. Of these individuals, 48% were female and 34% were over the age of 60 years old. Regarding race, interested individuals directly represented the intended use population: 17% were Black or African American, 2.7% were Asian, and 1.3% were Native Hawaiian, Pacific Islander, American Indian, or Alaskan Native. With respect to ethnicity, 8.4% identified as Hispanic or Latinx. The decentralized approach also permitted access to individuals with socioeconomic healthcare inequities: 27% had income under $29,999 and 14% were on Medicaid. Individuals were derived from all 48 continental United States, and of those who reported their residential location, approximately 3% were from rural areas. Conclusions: Use of a decentralized recruitment strategy permitted highly successful enrollment in the face of challenges associated with COVID-19. With respect to race, ethnicity, socioeconomic status, and geography, all metrics represented significantly more diverse populations than observed in traditional clinical studies. Decentralized enrollment mitigated selection bias, and will result in data more reflective of the intended use population. Clinical trial information: NCT04739722.

Jointly Predicting Future Content in Multiple Social Media Sites Based on Multi-task Learning

User-generated contents (UGC) in social media are the direct expression of users’ interests, preferences, and opinions. User behavior prediction based on UGC has increasingly been investigated in recent years. Compared to learning a person’s behavioral patterns in each social media site separately, jointly predicting user behavior in multiple social media sites and complementing each other (cross-site user behavior prediction) can be more accurate. However, cross-site user behavior prediction based on UGC is a challenging task due to the difficulty of cross-site data sampling, the complexity of UGC modeling, and uncertainty of knowledge sharing among different sites. For these problems, we propose a Cross-Site Multi-Task (CSMT) learning method to jointly predict user behavior in multiple social media sites. CSMT mainly derives from the hierarchical attention network and multi-task learning. Using this method, the UGC in each social media site can obtain fine-grained representations in terms of words, topics, posts, hashtags, and time slices as well as the relevances among them, and prediction tasks in different social media sites can be jointly implemented and complement each other. By utilizing two cross-site datasets sampled from Weibo, Douban, Facebook, and Twitter, we validate our method’s superiority on several classification metrics compared with existing related methods.

Psychological status of mass people in the capital city of Bangladesh during COVID-19: Do home quarantine challenges matter?

This study aimed to identify the impact of home quarantine challenges on the psychological status of mass people during Coronavirus Disease (COVID‐19) in the capital city (Dhaka) of Bangladesh. A highly organized questionnaire was created in “Google Form” and given to respondents at random via multiple social media sites, including WhatsApp, LinkedIn, and Messenger. However, a sample size of 208 people was taken and analysis was carried out using Statistical Package for the Social Sciences and Analysis of a Moment Structures software. Analysis showed that home quarantine challenges are positively related to the psychological status and home quarantine challenges have a significant impact on psychological status during COVID‐19. In Bangladesh, the major challenges to home quarantine are lack of consciousness, lack of medical support, and the possibility of losing their current jobs. This study provides insight into the challenges of home quarantine and the psychological status of Bangladeshi people. The findings may be helpful for policymakers in identifying necessary measures to overcome these challenges. Only Dhaka division was taken out of seven divisions in Bangladesh to conduct this study, which might be difficult to generalize the findings of this study. Thus, in the future, it is suggested that more respondents from other divisions need to be covered.

Digital start-up access to venture capital financing: What signals quality?

What signals quality for digital startups seeking access to external finance? Analyzing a unique dataset from India, we investigate the impact of traditional quality signals (e.g., the founders’ years of experience, prior industry experience); networking signals (e.g., degrees from elite educational institutes, the breadth of an entrepreneurial team); and digital signals (e.g., the social media presence of the startup on multiple social media sites) on the access to venture capital financing. We find that while networking and digital signals positively impact access to financing, traditional human capital signals do not have any significant impact on the process.

JUULing Epidemic Among Youth: A Guide to Devices, Terminology, and Interventions.

JUULing Epidemic Among Youth: A Guide to Devices, Terminology, and Interventions.

An experimental study on feature engineering and learning approaches for aggression detection in social media

With the widespread of modern technologies and social media networks, a new form of bullying occurring anytime and anywhere has emerged. This new phenomenon, known as cyberaggression or cyberbullying, refers to aggressive and intentional acts aiming at repeatedly causing harm to other person involving rude, insulting, offensive, teasing or demoralising comments through online social media. As these aggressions represent a threatening experience to Internet users, especially kids and teens who are still shaping their identities, social relations and well-being, it is crucial to understand how cyberbullying occurs to prevent it from escalating. Considering the massive information on the Web, the developing of intelligent techniques for automatically detecting harmful content is gaining importance, allowing the monitoring of large-scale social media and the early detection of unwanted and aggressive situations. Even though several approaches have been developed over the last few years based both on traditional and deep learning techniques, several concerns arise over the duplication of research and the difficulty of comparing results. Moreover, there is no agreement regarding neither which type of technique is better suited for the task, nor the type of features in which learning should be based. The goal of this work is to shed some light on the effects of learning paradigms and feature engineering approaches for detecting aggressions in social media texts. In this context, this work provides an evaluation of diverse traditional and deep learning techniques based on diverse sets of features, across multiple social media sites.

Social Media as Social Transition Machinery

Social media, and people's online self-presentations and social networks, add complexity to people's experiences managing changing identities during life transitions. I use gender transition as a case study to understand how people experience liminality on social media. I qualitatively analyzed data from transition blogs on Tumblr (n=240), a social media blogging site on which people document their gender transitions, and in-depth interviews with transgender bloggers (n=20). I apply ethnographer van Gennep's liminality framework to a social media context and contribute a new understanding of liminality by arguing that reconstructing one's online identity during life transitions is a rite of passage. During life transitions, people present multiple identities simultaneously on different social media sites that together comprise what I call social transition machinery. Social transition machinery describes the ways that, for people facing life transitions, multiple social media sites and networks o"en remain separate, yet work together to facilitate life transitions.

Social Media Use and Access to Digital Technology in US Young Adults in 2016.

BackgroundIn 2015, 90% of US young adults with Internet access used social media. Digital and social media are highly prevalent modalities through which young adults explore identity formation, and by extension, learn and transmit norms about health and risk behaviors during this developmental life stage.ObjectiveThe purpose of this study was to provide updated estimates of social media use from 2014 to 2016 and correlates of social media use and access to digital technology in data collected from a national sample of US young adults in 2016.MethodsYoung adult participants aged 18-24 years in Wave 7 (October 2014, N=1259) and Wave 9 (February 2016, N=989) of the Truth Initiative Young Adult Cohort Study were asked about use frequency for 11 social media sites and access to digital devices, in addition to sociodemographic characteristics. Regular use was defined as using a given social media site at least weekly. Weighted analyses estimated the prevalence of use of each social media site, overlap between regular use of specific sites, and correlates of using a greater number of social media sites regularly. Bivariate analyses identified sociodemographic correlates of access to specific digital devices.ResultsIn 2014, 89.42% (weighted n, 1126/1298) of young adults reported regular use of at least one social media site. This increased to 97.5% (weighted n, 965/989) of young adults in 2016. Among regular users of social media sites in 2016, the top five sites were Tumblr (85.5%), Vine (84.7%), Snapchat (81.7%), Instagram (80.7%), and LinkedIn (78.9%). Respondents reported regularly using an average of 7.6 social media sites, with 85% using 6 or more sites regularly. Overall, 87% of young adults reported access or use of a smartphone with Internet access, 74% a desktop or laptop computer with Internet access, 41% a tablet with Internet access, 29% a smart TV or video game console with Internet access, 11% a cell phone without Internet access, and 3% none of these. Access to all digital devices with Internet was lower in those reporting a lower subjective financial situation; there were also significant differences in access to specific digital devices with Internet by race, ethnicity, and education.ConclusionsThe high mean number of social media sites used regularly and the substantial overlap in use of multiple social media sites reflect the rapidly changing social media environment. Mobile devices are a primary channel for social media, and our study highlights disparities in access to digital technologies with Internet access among US young adults by race/ethnicity, education, and subjective financial status. Findings from this study may guide the development and implementation of future health interventions for young adults delivered via the Internet or social media sites.

“I will not hate myself because you cannot accept me”: Problematizing empowerment and gender-diverse selfies

ABSTRACTWhile selfies of beautiful cisgender women are declaimed by mainstream media as narcissistic and facile, some body-positive feminists and queer theorists argue that selfies can be empowering. They claim self-representation by traditionally stigmatized people can challenge normative presentations of beauty and gender. This article problematizes “empowerment” as a definitive and/or productive frame and argues instead for observation and analysis of “privilege” in situated practice. In this article I combine analysis of a collection of online cultural artifacts (including nonbinary selfies on Tumblr) and interviews with a small group of trans* social media storytellers to explore theoretical tensions between gender fluidity and identity fragmentation across multiple social media sites and practices. Gender-diverse digital self-representation encompasses both “consistent” androgyny, nonbinary, agender, and so on, and “emergent” presentations-in-flux. I assert that the ongoing iteration of self across social media—implied by self (re)presentation—can have simultaneous and contradictory political significance. I conclude that networked interpersonal complications frame understandings of empowerment, as perhaps they always have done.

The Creation of an Arabic Emotion Ontology Based on E-Motive

Abstract: There is an increased interest in social media monitoring to analyse massive, free form, short user-generated text from multiple social media sites such as Facebook, WhatsApp and Twitter. Companies are interested in sentiment analysis to understand customers’ opinions about their products/services. Governments and law enforcement agencies are interested in identifying threats to safeguard their country's national security. They are actively seeking ways to monitor and analyse the public's responses to various services, activities and events, especially since social media has become a valuable real-time resource of information. This study builds on prior work that focused on sentiment classification (i.e., positive, negative). This study primarily aims to design and develop a social sentiment-parsing algorithm for capturing and monitoring an extensive and comprehensive range of emotions from Arabic social media text. The study contributes to the field of sentiment analysis (opinion mining) and can subsequently be used for web mining, cleansing and analytics.

Relationship quality in higher education marketing: the role of social media engagement

ABSTRACTThe landscape in consumer marketing is changing due to the rise in popularity of social media. This shift has also affected how higher education institutions build relationships with their stakeholders. This study explores how social media engagement impacts relationship quality between the university and one of its key stakeholder groups, students. Data were collected via an online survey and analyzed using the Mann–Whitney U test, regression and the Kruskal–Wallis test. Results indicated a positive association between students following a university via social media and the perception of having a high-quality relationship with their university, and that following a university on multiple social media sites leads to an even higher perception of relationship quality. The results provide important and timely implications for both universities and higher education marketers. Our findings suggest that higher education marketers should invest resources in social media communications to form high-quality relationships with their stakeholders.

Walls-in-one: usage and temporal patterns in a social media aggregator.

The continual launches of new online social media that meet the most varied people’s needs are resulting in a simultaneous adoption of different social platforms. As a consequence people are pushed to handle their identity across multiple platforms. However, due the to specialization of the services, people’s identity and behavior are often partial, incomplete and scattered in different “places”. To overcome this identity fragmentation and to give an all-around picture of people’s online behavior, in this paper we perform a multidimensional analysis of users across multiple social media sites. Our study relies on a new rich dataset collecting information about how and when users post their favorite contents, about their centrality on different social media and about the choice of their username. Specifically we gathered the posting activities and social sites usage from Alternion, a social media aggregator.The analysis of social media usage shows that Alternion data reflect the novel trend of today’s users of branching out into different social platforms. However the novelty is the multidimensional and longitudinal nature of the dataset. Having at our disposal users’ degree in five different social networks, we performed a rank correlation analysis on users’ degree centrality and we find that the degrees of a given user are scarcely correlated. This is suggesting that the individuals’ importance changes from medium to medium. The longitudinal nature of the dataset has been exploited to investigate the posting activity. We find a slightly positive correlation on how often users publish on different social media and we confirm the burstiness of the posting activities extending it to multidimensional time-series. Finally we show that users tend to use similar usernames to keep their identifiability across social sites.