Search Results (7,266)

In recent years, the proliferation of fake news and misinformation has grown exponentially, far surpassing that of genuine news and posing a serious threat to social stability. Existing research in fake news detection primarily applies contrastive learning methods with a single-hot labeling strategy. The issue does not lie with contrastive learning as a technique but with its current application in fake news detection systems. Specifically, these systems penalize all negative samples equally due to the use of single-hot labeling, thus overlooking the underlying semantic relationships among negative samples. As a result, contrastive learning models tend to learn from simple samples while neglecting highly deceptive samples located at the boundary between true and false, as well as the heterogeneity of text-image features, which complicates cross-modal fusion. To mitigate these known limitations in current applications, this paper proposes a fake news detection method based on contrastive learning and cross-modal interaction. First, a consistency-aware soft-label contrastive learning mechanism based on semantic similarity is designed to provide more granular supervision signals for contrastive learning. Secondly, a difficult negative sample mining strategy based on a similarity matrix is designed to optimize the symmetry alignment of image and text features, which effectively improves the model’s ability to discriminate boundary samples. To further optimize the feature fusion process, a cross-modal interaction module is designed to learn the symmetric interaction relationship between image and text features. Finally, an attention mechanism is designed to adaptively adjust the contributions of text-image features and interaction features, forming the final multimodal feature representation. Experiments are conducted on two major social media platform datasets, and compared with existing methods, the proposed method effectively improves the detection capability of fake news. Full article

Background and Objectives: With the widespread integration of digital media into daily life, social media addiction (SMA) has become a growing concern for university students’ mental health. Based on attachment theory, this study examined how attachment anxiety and avoidance influence SMA through fear of negative evaluation (FNE) and loneliness. Methods: A sample of 400 Chinese university students completed the 16-item short version of the Experiences in Close Relationships Scale (ECR), the 8-item Brief Fear of Negative Evaluation Scale (BFNE), the 6-item Revised UCLA Loneliness Scale–Short Form (RULS-6), and the 6-item Bergen Social Media Addiction Scale (BSMAS). Using the PROCESS macro (Model 6), a chained mediation model was tested. Results: Attachment anxiety positively predicts SMA (β = 0.42); the chained mediation pathway through FNE and loneliness accounts for ab = 0.06 of this effect, alongside additional single-mediator paths. In contrast, attachment avoidance shows a weaker total effect (β = −0.08) and a small negative chained mediation effect (ab = −0.02), offset by opposing single-mediator paths via FNE (negative) and loneliness (positive), resulting in a nonsignificant total indirect effect. Discussion: These findings suggest that in the Chinese cultural context, where social evaluation and belonging are emphasized, insecure attachment may heighten emotional reliance on social media. This study elucidates the socio-emotional mechanisms underlying SMA and extends the application of attachment theory to the digital media environment. Full article

This study focuses on cultural and urban development policies under China’s 14th Five-Year Plan, exploring the content and semantic structure of discussions on the “East Asian Cultural Capital” project on the Weibo platform. It analyzes how national cultural development policies are reflected in the discourse system related to the “East Asian Cultural Capital” on social media and emphasizes the guiding role of policies in the dissemination of online culture. When China announced the 14th Five-Year Plan in 2021, the strategic direction and policy framework for cultural development over the five-year period from 2021 to 2025 were clearly outlined. This study employs text mining and semantic network analysis methods to analyze user-generated content on Weibo from 2023 to 2024, aiming to understand public perception and discourse trends. Word frequency and TF-IDF analyses identify key terms and issues, while centrality and CONCOR clustering analyses reveal the semantic structure and discourse communities. MR-QAP regression is employed to compare network changes across the two years. Findings highlight that urban cultural development, heritage preservation, and regional exchange are central themes, with digital media, cultural branding, trilateral cooperation, and cultural–economic integration emerging as key factors in regional collaboration. Full article

This study examines the influence of mass media on the career choices of high school students from Brașov County, Romania, with a focus on their underlying motivational factors. Employing a quantitative design, it draws on data from a standardized questionnaire completed by 1314 students from local high schools. Descriptive and inferential statistical methods were used in the analysis. While some students identify mass media as a key source of career guidance, documentaries and career fairs are more frequently cited as trusted sources. Students’ perceptions of mass media are ambivalent: 55.1% see it as manipulative, while 41.7% and 24.7% acknowledge its informative and educational roles. Personal motivation emerges as the most significant influence, with 64.8% guided by individual talents and abilities, compared to a lower influence from family or media role models. Correlational analysis highlights the importance of personal development, creativity, and collaboration in career motivation. This study suggests that mass media indirectly shapes students’ aspirations by reinforcing values like social recognition, mobility, and identity. Finally, it reveals a strong link between career interest and expectations for respectful, stable, and growth-oriented work environments, pointing to a pragmatic orientation toward professional sustainability. Full article

The proliferation of social media has paradoxically facilitated the widespread dissemination of fake news, impacting individuals, politics, economics, and society as a whole. Despite the increasing scholarly research on this phenomenon, a significant gap exists regarding its dynamics in developing countries, particularly how predictors of fake news sharing interact, rather than merely their net effects. To acquire a more nuanced understanding of fake news sharing behavior, we propose identifying the direct and complex interplay among key variables by utilizing a dual analytical framework, leveraging Structural Equation Modeling (SEM) for linear relationships and Fuzzy-Set Qualitative Comparative Analysis (fsQCA) to uncover asymmetric patterns. Specifically, we investigate the influence of news-find-me orientation, social media trust, information-sharing tendencies, and status-seeking motivation on the propensity of fake news sharing behavior. Additionally, we delve into the moderating influence of social media literacy on these observed effects. Based on a cross-sectional survey of 1028 Haitian social media users, the SEM analysis revealed that news-find-me perception had a negative but statistically insignificant influence on fake news sharing behavior. In contrast, information sharing exhibited a significant negative association. Trust in social media was positively and significantly linked to fake news sharing behavior. Meanwhile, status-seeking motivation was positively associated with fake news sharing behavior, although the association did not reach statistical significance. Crucially, social media literacy moderated the effects of trust and information sharing. Interestingly, fsQCA identified three core configurations for fake news sharing: (1) low status seeking, (2) low information-sharing tendencies, and (3) a unique interaction of low “news-find-me” orientation and high social media trust. Furthermore, low social media literacy emerged as a direct core configuration. These findings support the urgent need to prioritize social media literacy as a key intervention in combating the dissemination of fake news. Full article

Existing approaches to social media sentiment analysis typically focus on static classification, offering limited foresight into how public opinion evolves. This study addresses that gap by introducing the Multi-Feature Sentiment-Driven Forecasting (MFSF) framework, a novel pipeline that enhances sentiment trend prediction by integrating rich contextual information from text. Using state-of-the-art transformer models on the Sentiment140 dataset, our framework extracts three concurrent signals from each tweet: sentiment polarity, aspect-based scores (e.g., ‘price’ and ‘service’), and topic embeddings. These features are aggregated into a daily multivariate time series. We then employ a SARIMAX model to forecast future sentiment, using the extracted aspect and topic data as predictive exogenous variables. Our results, validated on the historical Sentiment140 Twitter dataset, demonstrate the framework’s superior performance. The proposed multivariate model achieved a 26.6% improvement in forecasting accuracy (RMSE) over a traditional univariate ARIMA baseline. The analysis confirmed that conversational aspects like ‘service’ and ‘quality’ are statistically significant predictors of future sentiment. By leveraging the contextual drivers of conversation, the MFSF framework provides a more accurate and interpretable tool for businesses and policymakers to proactively monitor and anticipate shifts in public opinion. Full article

Smartphones have revolutionized our daily lives, becoming portable pocket computers with easy internet access. India, the second-highest smartphone and internet user, experienced a significant rise in smartphone usage between 2013 and 2024. Prolonged smartphone use, exceeding 20 min at a time, can lead to physical and mental health issues, including psychophysiological disorders. Digital devices and their extended exposure to blue light cause digital eyestrain, sleep disorders and visual-related problems. This research examines the impact of 1 h smartphone usage on visual fatigue among young Indian adults. A portable, low-cost system has been developed to measure visual activity to address this. The developed visual activity measurement system measures blink rate, inter-blink interval, and pupil diameter. Measured eye activity was recorded during 1 h smartphone usage of e-book reading, video watching, and social-media reels (short videos). Social media reels show increased screen variations, affecting pupil dilation and reducing blink rate due to continuous screen brightness and intensity changes. This reduction in blink rate and increase in inter-blink interval or pupil dilation could lead to visual fatigue. Full article

Physics concepts are considered an essential component of STEM (science, technology, engineering, and mathematics) education and fundamental for economic and technological development in the world. However, there can be student academic underperformance, such as the school environment, learning media and infrastructure, student interest and emotions, as well as social and economic development factors in communities. These problems are even more acute in rural areas of developing countries, where poverty is high and teachers often lack the necessary technological skills. The aim of this study was to evaluate the impact of a low-cost STEM tool focused on motivation in learning, in terms of five variables of interest in physics in rural areas, as well as the durability of the tools used to learn 12 physics concepts. A quasi-experimental study was conducted with the participation of 78 high school students, with an average age of 15.82 years, in a rural area of Guerrero, Mexico. The results showed that using the STEM tool significantly increased students’ interest in learning methodology, active participation, and attitude towards physics, facilitating the teacher’s work. In addition, the 3D construction kit used in the experimentation, besides being low-cost, proved to be affordable and durable, making it ideal for use in rural areas. Full article

Floods pose a growing threat globally, causing tragic loss of life, billions in economic damage annually, and disproportionately affecting socio-economically vulnerable populations. This paper aims to improve flood-risk communication for community leaders by exploring the application of artificial intelligence. We categorize U.S. flood information sources, review communication modalities and channels, synthesize the literature on community leaders’ roles in risk communication, and analyze existing technological tools. Our analysis reveals three key challenges: the fragmentation of flood information, information overload that impedes decision-making, and the absence of a unified communication platform to address these issues. We find that AI techniques can organize data and significantly enhance communication effectiveness, particularly when delivered through infographics and social media channels. Based on these findings, we propose FLAI (Flood Language AI), an AI-driven flood communication platform that unifies fragmented flood data sources. FLAI employs knowledge graphs to structure fragmented data sources and utilizes a retrieval-augmented generation (RAG) framework to enable large language models (LLMs) to produce contextualized narratives, including infographics, maps, and cost–benefit analyses. Beyond flood management, FLAI’s framework demonstrates how AI can transform public service data management and institutional AI readiness. By centralizing and organizing information, FLAI can significantly reduce the cognitive burden on community leaders, helping them communicate timely, actionable insights to save lives and build flood resilience. Full article

Fuel price increases have long been a contentious issue in Indonesia, sparking intense public and political debates. This study examines how digital media, particularly Kompas.com and Tempo.co, shape public discourse on fuel price hikes through mediatization. Using discourse network analysis, this study compares the political narratives surrounding fuel price increases during the administrations of Susilo Bambang Yudhoyono (2013) and Joko Widodo (2022). The findings reveal a shift in dominant discourse—opposition to price hikes was prominent in both periods, with government authority and economic justification emphasized in 2013, whereas concerns over rising living costs and social unrest dominated in 2022. This study highlights how mediatization has transformed policymaking from deliberative discussions into fragmented media battles, where digital platforms amplify competing narratives rather than facilitating consensus. Kompas.com predominantly featured counter-discourses, while Tempo.co exhibited stronger pro-government narratives in 2013. This study suggests that while digital media plays a crucial role in shaping policy perceptions, it does not necessarily translate into policy influence. It contributes to the broader understanding of the media’s role in policy debates. It underscores the need for more strategic government communication to manage public expectations and mitigate political unrest surrounding fuel price adjustments. Full article

With the popularization of social media, information related to public health events has seen explosive growth online, making it essential to accurately identify informative tweets with decision-making and management value for public health emergency response and risk monitoring. However, existing methods often suffer performance degradation during cross-event transfer due to differences in data distribution, and research specifically targeting public health events remains limited. To address this, we propose the Label Confusion Domain Adversarial Network (LCDAN), which innovatively integrates label confusion with domain adaptation to enhance the detection of informative tweets across different public health events. First, LCDAN employs an adversarial domain adaptation model to learn cross-domain feature representation. Second, it dynamically evaluates the importance of different source domain samples to the target domain through label confusion to optimize the migration effect. Experiments were conducted on datasets related to COVID-19, Ebola disease, and Middle East Respiratory Syndrome public health events. The results demonstrate that LCDAN significantly outperforms existing methods across all tasks. This research provides an effective tool for information detection during public health emergencies, with substantial theoretical and practical implications. Full article

In the attention-driven environment of online video platforms, understanding the factors that influence content selection and channel subscriptions is crucial for creators, marketers, and platform managers. This study investigates how thumbnails, view counts, video length, genre, and the number of advertisements affect user decision-making on YouTube. Grounded in random utility theory and reference-dependent preference theory, this study conducted a choice experiment with 525 respondents and employed a combined model of rank-ordered and binary logit methods to analyze viewing and subscription behaviors. The results indicate a significant preference for thumbnails with subtitles and shorter videos. Notably, we found evidence of reference-dependent effects, whereby a higher-than-expected number of ads decreased viewing probability, while a lower-than-expected number significantly increased subscription probability. This study advances our understanding of the factors that influence user behavior on social media, specifically in terms of viewing and subscribing, and empirically supports prospect theory in the online advertising market. Our findings offer both theoretical and practical insights into optimizing video content and monetization strategies in competitive social media markets. Full article

Understanding the social drivers of pro-environmental behavior in urban forests and green spaces is critical for addressing sustainability challenges. Subjective norms serve as a key pathway through which social expectations influence individuals’ behavioral intentions. Despite mixed findings in the literature regarding the impact of subjective norms on individuals’ intentions, there is a research gap about the determinants of this construct. This study was conducted to explore how social expectations shape perceived subjective norms among visitors of urban forests. A theoretical model was developed with subjective norms at its center, incorporating their predictors including social identity, media influence, interpersonal influence, and institutional trust, personal norms as a mediator, and behavioral intention as the outcome variable. Using structural equation modeling, data was collected and analyzed from a sample of visitors of urban forests in Tehran, Iran. The results revealed that subjective norms play a central mediating role in linking external social factors to behavioral intention. Social identity emerged as the strongest predictor of subjective norms, followed by media and interpersonal influence, while institutional trust had no significant effect. Subjective norms significantly influenced both personal norms and intentions, and personal norms also directly predicted intention. The model explained 50.9% of the variance in subjective norms and 39.0% in behavioral intention, highlighting its relatively high explanatory power. These findings underscore the importance of social context and internalized norms in shaping sustainable behavior. Policy and managerial implications suggest that strategies should prioritize community-based identity reinforcement, media engagement, and peer influence over top-down institutional messaging. This study contributes to environmental psychology and the behavior change literature by offering an integrated, empirically validated model. It also provides practical guidance for designing interventions that target both social and moral dimensions of environmental action. Full article

Social media platforms have emerged as valuable sources for mental health research, enabling the detection of conditions such as depression through analyses of user-generated posts. This manuscript offers practical, step-by-step guidance for applying machine learning and deep learning methods to mental health detection on social media. Key topics include strategies for handling heterogeneous and imbalanced datasets, advanced text preprocessing, robust model evaluation, and the use of appropriate metrics beyond accuracy. Real-world examples illustrate each stage of the process, and an emphasis is placed on transparency, reproducibility, and ethical best practices. While the present work focuses on text-based analysis, we discuss the limitations of this approach—including label inconsistency and a lack of clinical validation—and highlight the need for future research to integrate multimodal signals and gold-standard psychometric assessments. By sharing these frameworks and lessons, this manuscript aims to support the development of more reliable, generalizable, and ethically responsible models for mental health detection and early intervention. Full article

This study examines how Environmental, Social, and Governance (ESG) reporting is communicated and perceived on YouTube. A dataset of 553 relevant videos and 5060 user comments was extracted on 2 April 2025 ranging between 2014 and 2025, and sentiment, topic, and stance analyses were applied to both transcripts and comments. The majority of video content strongly endorsed ESG reporting, emphasizing themes such as transparency, regulatory compliance, and financial performance. In contrast, viewer comments revealed diverse stances, including skepticism about methodological inconsistencies, accusations of greenwashing, and concerns over politicization. Notably, statistical analysis showed minimal correlation between video sentiment and audience sentiment, suggesting that user perceptions are shaped by factors beyond the tone of the videos themselves. These findings underscore the need for more rigorous ESG frameworks, enhanced standardization, and proactive stakeholder engagement strategies. The study highlights the value of online platforms for capturing stakeholder feedback in real time, offering practical insights for organizations and policymakers seeking to strengthen ESG disclosure and communication. Full article