AMPLE: Emotion-Aware Multimodal Fusion Prompt Learning for Fake News Detection

Lecture notes in computer science, Год журнала: 2025, Номер unknown, С. 86 - 100

Опубликована: Янв. 1, 2025

Язык: Английский

---

Misinformation detection: datasets, models and performance

Online Information Review, Год журнала: 2025, Номер unknown

Опубликована: Янв. 3, 2025

Purpose This paper aims to understand the characteristics of current misinformation detection studies, including datasets used by researchers, computational models or algorithms being developed applied, and performance algorithms. Design/methodology/approach We first identified articles from Scopus database with inclusion exclusion criteria. Then a coding scheme was derived based on research questions. Next, datasets, models, were coded. The concluded answers questions future directions. Findings From 115 relevant published during 2019–2023 detection. found that most studies previously existing datasets. Twitter (now X) has been widely source for collecting social media data. ten frequently are identified. Most (96.1%) applied machine learning, especially deep learning models. advanced could achieve pretty high performance. For example, among 104 reporting accuracy, 44.2% achieved an accuracy 0.95 higher, 24.0% 0.90–0.94 accuracy. Research limitations/implications Our study only reviewed English included in database. Articles not reviewed. Practical implications indicates should be able detect if they willing do it. However, no system algorithm 100% Due complexity misinformation, users still need improve their capabilities evaluating information Internet. Social provides evidence policymakers platforms have capability detecting posted. These responsible alerting suspicious postings misinformation. Originality/value identifies computer research. findings will help companies, scientists, designers systems. It also students science latest Information professionals may work scientists

Язык: Английский

---

NLP-Driven Analysis of Users’ Reaction for Estimation of Information Disorder Propagation

Lecture notes on data engineering and communications technologies, Год журнала: 2025, Номер unknown, С. 417 - 425

Опубликована: Янв. 1, 2025

Язык: Английский

---

Tropes and the EmotAix lexicon for evaluating the emotional tonality of French verbal association corpora in social representation studies

Language Resources and Evaluation, Год журнала: 2025, Номер unknown

Опубликована: Апрель 28, 2025

Язык: Английский

---

A Theory-Driven Approach to Fake News/Information Disorder Analysis and Explanation via Target-Based Emotion–Stance Analysis (TESA) and Interpretive Graph Generation (IGG)

Social Science Computer Review, Год журнала: 2025, Номер unknown

Опубликована: Май 2, 2025

Information disorder (IDO) presents a persistent challenge to society, necessitating innovative approaches understanding its dynamics beyond just merely detecting it. This study introduces theory-driven framework that integrates advanced natural language processing (NLP) with deep learning, utilizing the target-based emotion–stance analysis (TESA) approach analyze emotion and stance within IDO content. Complementing TESA, interactive graph generation (IGG) is applied for scalable interpretable qualitative analyses. Employing mixed-methods approach, leverages TESA target-centric analysis, evaluating classifiers on both human-annotated synthetic datasets. Additionally, explores data using generative AI enrich applying IGG map complex interactions. The also found integrating developed from human annotations enhanced model performance, particularly classification tasks. Results demonstrate narratives significantly differ non-IDO narratives, frequently leveraging negative emotions such as anger disgust manipulate public perception. proved effective in capturing these nuanced variations, while facilitated triangulation of findings via interpretation emotional revealing content often amplifies polarizing antagonistic perspectives. By combining IGG, this research emphasizes importance NLP extract examine nuances toward targets interest context. not only deepens theoretical insights into IDO’s persuasive mechanisms but supports development practical tools analyzing managing influence discourse.

Язык: Английский

---

Methodology for visual analysis of psychological tension in online discourse on coronavirus vaccination

Global Knowledge Memory and Communication, Год журнала: 2025, Номер unknown

Опубликована: Апрель 26, 2025

Purpose This study aims to introduce a novel methodology for visually analyzing psychological tension in social networks, particularly the context of disturbances related coronavirus vaccination. It also enhance interpretation online discourse dynamics by integrating mathematical, linguistic and visual-analytical methods. Design/methodology/approach The uses comprehensive approach, including tweet array generation via Vicinitas API, key term extraction, sentiment analysis visualization tools such as Word-Cloud, VosViewer, Gephisto Gephi. is tested on protests against vaccination evaluate its effectiveness capturing intricacies digital discourse. Findings identifies themes, stress linked vaccination, protest movements sentiments regarding trust belief. Hierarchical representations visualizations reveal nuances within discourse, demonstrating methodology’s capacity discern patterns media interactions. Practical implications Practically, this offers robust tool monitoring interpreting scenarios demanding immediate response, public health crises. Public officials can use method detect early signs misinformation or distress, enabling timely interventions. Policymakers analysts leverage these insights design communication strategies that build mitigate anxiety, aligning with societal needs transparency accountability. Originality/value research introduces integration computational expert insights, specifically designed real-time Unlike previous methods focus text evaluation independently, approach uniquely combines predictive modeling, offering lens understanding interactions during sensitive

Язык: Английский

---

Enhanced Sentiment Intensity Regression Through LoRA Fine-Tuning on Llama 3

IEEE Access, Год журнала: 2024, Номер 12, С. 108072 - 108087

Опубликована: Янв. 1, 2024

Sentiment analysis and emotion detection are critical research areas in natural language processing (NLP), offering benefits to numerous downstream tasks. Despite the widespread application of pre-trained models large (LLMs) sentiment analysis, most previous works have focused on polarity or classification, neglecting finer-grained task intensity regression, which prevents precise capture hindering model performance complex scenarios diverse applications. To address this issue, we enhance Roberta with an efficient additive attention mechanism adaptive weighted Huber loss function, notably improving its regression. Based SemEval 2017 2018 datasets, employ prompt engineering construct fine-tuned further enriched outputs from enhanced model. We then fine-tune Llama 3 using Low-Rank Adaptation (LoRA) within Unsloth framework. Experimental results demonstrate that our RoBERTa significantly outperforms baseline models. Furthermore, LoRA 3-8B other LLMs similar parameter scales. Our method improves MAE by 0.015 MSE 0.0054 dataset, achieving a Pearson correlation coefficient 0.8441. On it 0.0416 0.043, increased 0.8268, demonstrates superior predictive power robustness approach.

Язык: Английский

---

TCMT: Target-oriented Cross Modal Transformer for multimodal aspect-based sentiment analysis

Expert Systems with Applications, Год журнала: 2024, Номер unknown, С. 125818 - 125818

Опубликована: Ноя. 1, 2024

Язык: Английский

---

Detecting fake news on social networks via linguistic features and information-seeking patterns during the Covid-19 period

Annals of Operations Research, Год журнала: 2024, Номер unknown

Опубликована: Ноя. 26, 2024

Язык: Английский

---

REDAffectiveLM: leveraging affect enriched embedding and transformer-based neural language model for readers’ emotion detection

Knowledge and Information Systems, Год журнала: 2024, Номер 66(12), С. 7495 - 7525

Опубликована: Авг. 19, 2024

Язык: Английский

---