A5023817571- Hate Speech and Cyberbullying Detection
- Social Media and Politics
- Misinformation and Its Impacts
- Complex Network Analysis Techniques
- Opinion Dynamics and Social Influence
- Digital Marketing and Social Media
- Topic Modeling
- Spam and Phishing Detection
- Sentiment Analysis and Opinion Mining
- Natural Language Processing Techniques
- Big Data and Business Intelligence
- Media Influence and Politics
- Advanced Malware Detection Techniques
- Digital Transformation in Industry
- Knowledge Management and Sharing
- Public Relations and Crisis Communication
- Computational and Text Analysis Methods
- Wikis in Education and Collaboration
- Team Dynamics and Performance
- AI in Service Interactions
- Media Studies and Communication
- Research Data Management Practices
- Data Quality and Management
- Cybercrime and Law Enforcement Studies
- Impact of Technology on Adolescents
University of Edinburgh
2020-2025
University of Duisburg-Essen
2017-2023
Tokyo Institute of Technology
2023
Administration for Community Living
2023
IT University of Copenhagen
2023
American Jewish Committee
2023
Network Technologies (United States)
2020
Bell (Canada)
2002
Since an ever-increasing part of the population makes use social media in their day-to-day lives, data is being analysed many different disciplines. The analytics process involves four distinct steps, discovery, collection, preparation, and analysis. While there a great deal literature on challenges difficulties involving specific analysis methods, hardly exists research stages preparation. To address this gap, we conducted extended structured through which identified addressed solutions...
Some users of social media are spreading racist, sexist, and otherwise hateful content. For the purpose training a hate speech detection system, reliability annotations is crucial, but there no universally agreed-upon definition. We collected potentially messages asked two groups internet to determine whether they were or not, should be banned not rate their degree offensiveness. One was shown definition prior completing survey. aimed assess can annotated reliably, extent which existing...
In this article, we introduce and discuss the pervasive issue of bias in large language models that are currently at core mainstream approaches to Natural Language Processing (NLP). We first data selection bias, is, caused by choice texts make up a training corpus. Then, survey different types social evidenced text generated trained on such corpora, ranging from gender age, sexual orientation ethnicity, religion culture. conclude with directions focused measuring, reducing, tackling...
So-called 'social bots' have garnered a lot of attention lately. Previous research showed that they attempted to influence political events such as the Brexit referendum and US presidential elections. It remains, however, somewhat unclear what exactly can be understood by term bot'. This paper addresses need better understand intentions bots on social media develop shared understanding how 'social' differ from other types bots. We thus describe systematic review publications researched bot...
Abstract Organizations introduce virtual assistants (VAs) to support employees with work-related tasks. VAs can increase the success of teamwork and thus become an integral part daily work life. However, effect on teams remains unclear. While social identity theory describes identification team members continued existence a group identity, concept extended self refers incorporation possessions into one’s sense self. This raises question which approach applies as teammates. The article...
Widespread mis- and disinformation during the COVID-19 social media "infodemic" challenge effective response of Emergency Management Agencies (EMAs). Conversational Agents (CAs) have potential to amplify distribute trustworthy information from EMAs general public in times uncertainty. However, structure responsibilities such are different comparison traditional commercial organizations. Consequently, Information Systems (IS) design approaches for CAs not directly transferable this type...
Abstract When attempting to solve a problem, humans call upon cognitive resources. These resources are limited, and the degree of their utilisation is described as load. While number parameters be taken into account processed by modern-day knowledge workers increases, do not. Research shows that too high load can increase stress failure rates decrease work satisfaction performance employees. It thus in interest organisations reduce employees keep it at moderate level. One way achieve this...
Modern computational text classification methods have brought social scientists tantalizingly close to the goal of unlocking vast insights buried in data—from centuries historical documents streams media posts. Yet three barriers still stand way: tedious labor manual annotation, technical complexity that keeps these tools out reach for many researchers, and, perhaps most critically, challenge bridging gap between sophisticated algorithms and deep theoretical understanding already developed...
During a crisis, authorities need to effectively disseminate information. We address the problem of deciding how crisis-related information should be published on Facebook reach as many people possible. examine three recent terrorist attacks in Berlin, London and Stockholm. Our specific focus lies with official pages by municipalities emergency service agencies. collected posts about events, along number shares, likes emotional reactions them. In regression analysis, several variables were...
As false information may spread rapidly on social media, a profound understanding of how it can be debunked is required. This study offers empirical insights into the development rumors after they are debunked, various user groups who involved in process, and their network structures. crisis situations highly sensitive to rumors, Twitter posts from during 2017 G20 summit examined. Tweets regarding five that were this event manually coded following categories: rumor, debunking message,...
Computational social science uses computational and statistical methods in order to evaluate interaction. The public availability of data sets is thus a necessary precondition for reliable replicable research. These allow researchers benchmark the they develop, test generalizability their findings, build confidence results. When media are concerned, sharing often restricted legal or privacy reasons, which makes comparison replicability research results infeasible. Social analytics research,...
Automated sentiment analysis can help efficiently detect trends in patients' moods, consumer preferences, political attitudes and more. Unfortunately, like many natural language processing techniques, show bias against marginalised groups. We illustrate this point by showing how six popular tools respond to sentences about queer identities, expanding on existing work gender, ethnicity disability. find evidence of several including the two models from Google Amazon that seem have been subject...
Research data management (RDM) is an important prerequisite for a substantial and sustainable contribution to knowledge. There pressing need examine why researchers hesitate store, annotate, share manage their research data. To model underlying psychological factors influencing researchers' refusal conduct RDM, the social exchange theory extended with elements from prospect theory. Thus, it allows insights into decision-making, illustrates role of cost benefit evaluations under uncertainty....