Automated stance detection in complex topics and small languages: The challenging case of immigration in polarizing news media

PLoS ONE, Journal Year: 2024, Volume and Issue: 19(4), P. e0302380 - e0302380

Published: April 26, 2024

Automated stance detection and related machine learning methods can provide useful insights for media monitoring academic research. Many of these approaches require annotated training datasets, which limits their applicability languages where may not be readily available. This paper explores the large language models automated in a challenging scenario, involving morphologically complex, lower-resource language, socio-culturally complex topic, immigration. If approach works this case, it expected to perform as well or better less demanding scenarios. We annotate set pro- anti-immigration examples train compare performance multiple models. also probe usability GPT-3.5 (that powers ChatGPT) an instructable zero-shot classifier same task. The supervised achieve acceptable performance, but yields similar accuracy. As latter does tuning with data, constitutes potentially simpler cheaper alternative text classification tasks, including languages. further use best-performing model investigate diachronic trends over seven years two corpora Estonian mainstream right-wing populist news sources, demonstrating analytics settings even scenarios, discuss correspondences between changes real-world events.

Language: Английский

---

Polarization on social media: Comparing the dynamics of interaction networks and language‐based opinion distributions

Political Psychology, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 10, 2025

Abstract When people share information and converse on social media, they create “echo chambers” through preferential attachment to like‐minded opinions already support. A great deal of research uses interactional ties between people—created by retweeting following—to identify study polarization in networks. Some this work then language analysis characterize the concerns subcommunities network. We used machine learning “speaker landscapes” that can user (in tweets about COVID‐19 vaccination) independently networks created interactions via retweeting. In contrast prevailing assumptions, we found distances users interaction did not predict their similarity very well. compared effect a polarizing event (the declaration pandemic) communities retweet speaker landscapes. was done both support criticize claims Democrats Republicans emerged much more strongly landscapes than The results suggest different cognitive‐motivational dynamics affect who interact with what say raising questions how is promote polarization.

Language: Английский

---

A systematic review of automated hyperpartisan news detection

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(2), P. e0316989 - e0316989

Published: Feb. 21, 2025

Hyperpartisan news consists of articles with strong biases that support specific political parties. The spread such increases polarization among readers, which threatens social unity and democratic stability. Automated tools can help identify hyperpartisan in the daily flood articles, offering a way to tackle these problems. With recent advances machine learning deep learning, there are now more methods available address this issue. This literature review collects organizes different used previous studies on detection. Using PRISMA methodology, we reviewed systematized approaches datasets from 81 published January 2015 2024. Our analysis includes several steps: differentiating detection similar tasks, identifying text sources, labeling methods, evaluating models. We found some key gaps: is no clear definition hyperpartisanship Computer Science, most English, highlighting need for minority languages. Moreover, tendency models perform better than traditional but Large Language Models’ (LLMs) capacities domain have been limitedly studied. paper first systematically detection, laying solid groundwork future research.

Language: Английский

---

Machine-assisted quantitizing designs: augmenting humanities and social sciences with artificial intelligence

Humanities and Social Sciences Communications, Journal Year: 2025, Volume and Issue: 12(1)

Published: Feb. 28, 2025

Language: Английский

---

The role of English language profiency on political socialization and democracy

Cogent Arts and Humanities, Journal Year: 2025, Volume and Issue: 12(1)

Published: March 20, 2025

Language: Английский

---

Beyond Buzzwords: The Development of Large Language Models and Their Use in Advertising and Strategic Communication Research

Journal of Current Issues & Research in Advertising, Journal Year: 2025, Volume and Issue: unknown, P. 1 - 40

Published: May 19, 2025

Language: Английский

---

Analyzing “Jayu” in South Korean presidential rhetoric: a comprehensive study from 1948–2023 with a focus on the Yoon Suk Yeol administration

Humanities and Social Sciences Communications, Journal Year: 2024, Volume and Issue: 11(1)

Published: June 8, 2024

Abstract The current study examines the strategic use of “Jayu” (freedom or liberty) in South Korean politics, with a focus on President Yoon Suk Yeol’s administration, where it symbolizes conservative ideologies and political identity. Employing Natural Language Processing, time-series analysis, visualization techniques, research analyzes presidential speeches to explore Yoon’s marked emphasis Jayu, indicative strong allegiance. findings reveal significant association between utilization Jayu strategies, underscoring its crucial role strategy function garnering support from factions within polarized context. discourse, characterized by an extensive fosters polarization partisanship, moving away inclusive dialog. This illuminates symbolic language communication identity formation, providing insights into interplay rhetoric ideological positions intricate landscape Korea.

Language: Английский

---

Natural Affect Detection (NADE): Inferring Emotional Expression From Text Through Emojis

SSRN Electronic Journal, Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 1, 2024

Brand perceptions are increasingly influenced by consumers sharing their opinions online, with rich emotions conveyed in texts playing a key role shaping consumer decisions. Consequently, brand managers and researchers require tools to accurately extract fine-grained from social media texts. Despite the central of communications, existing for such detailed emotional analysis have several drawbacks, which is why most marketing papers monitoring still use sentiment when extracting cues text. Applications beyond pure typically rely on dictionary-based methods, despite limited vocabulary. Advanced machine learning models address this challenge but extensive computing resources programming skills. Recent large language associated financial environmental costs. To these issues, paper introduces Nade, text-to-emoji-to-emotion converter that first "emojifies" natural then transforms obtained emojis into intensity measures well-studied theory-grounded emotions. Our robust, adaptable, cost-efficient, resource-efficient, easy-to-use via an online app packages Python R. Using human raters state-of-the-art converters as benchmarks, validates Nade illustrates how it applications data various platforms.

Language: Английский

---

Political orientation in media treatment of police violence: Evidence from modal adjectives

Discourse & Society, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 12, 2024

The goal of this paper is to unveil possible correlations between the political orientation newspapers and their treatment police violence. We consider three different news publications with diverging orientations, namely Jacobin (Left), Breitbart (Right), New York Times (Center). performed a corpus study that relies on two categorizations: new ontology for violence situations, identifying set recurrent themes, use distributions modal adjectives across these themes as revealing stances toward them. Modal are highly polysemous, our analysis distinguishes epistemic readings relating factual truth, evaluative norms. Our shows left right leaning journals share similar uses but differ in adjectives. These results could suggest responsible differences lie stance adopt.

Language: Английский

---