Search

The paper will investigate the complexities involved in moderating casteist content on major social media platforms. Focusing on the challenges of categorizing hate speech targeting India's marginalized caste-oppressed minorities, the research aims to delve into the mechanisms utilized by platforms for content moderation. It will critically analyze existing categorization systems, exploring the efficacy and limitations in identifying and addressing casteist comments. By leveraging information sciences concepts like grounded coding and taxonomies, the study will elucidate the complexities of classifying caste-related hate speech within these systems. Moreover, it will highlight the implications of inadequate categorization on content visibility and community impact.

BackgroundManually analyzing public health-related content from social media provides valuable insights into the beliefs, attitudes, and behaviors of individuals, shedding light on trends and patterns that can inform public understanding, policy decisions, targeted interventions, and communication strategies. Unfortunately, the time and effort needed from well-trained human subject matter experts makes extensive manual social media listening unfeasible. Generative large language models (LLMs) can potentially summarize and interpret large amounts of text, but it is unclear to what extent LLMs can glean subtle health-related meanings in large sets of social media posts and reasonably report health-related themes.ObjectiveWe aimed to assess the feasibility of using LLMs for topic model selection or inductive thematic analysis of large contents of social media posts by attempting to answer the following question: Can LLMs conduct topic model selection and inductive thematic analysis as effectively as humans did in a prior manual study, or at least reasonably, as judged by subject matter experts?MethodsWe asked the same research question and used the same set of social media content for both the LLM selection of relevant topics and the LLM analysis of themes as was conducted manually in a published study about vaccine rhetoric. We used the results from that study as background for this LLM experiment by comparing the results from the prior manual human analyses with the analyses from 3 LLMs: GPT4-32K, Claude-instant-100K, and Claude-2-100K. We also assessed if multiple LLMs had equivalent ability and assessed the consistency of repeated analysis from each LLM.ResultsThe LLMs generally gave high rankings to the topics chosen previously by humans as most relevant. We reject a null hypothesis (P

Climate communication scientists search for effective message strategies to engage the ambivalent public in support of climate advocacy. The personal experience of wildfire is expected to render climate change impacts more concretely, pointing to a potential message strategy to engage the public. This study examined Twitter discourse related to climate change during the onset of 20 wildfires in California between the years 2017 and 2021. In this mixed method study, we analyzed tweets geographically and temporally proximal to the occurrence of wildfires to discover framings and examined how frequencies in climate framings changed before and after fires. Results identified three predominant climate framings: linking wildfire to climate change, suggesting climate actions, and attributing climate change to adversities besides wildfires. Mean tweet frequencies linking wildfire to climate change and attributing adversities increased significantly after the onset of fire. While suggesting climate action tweets also increased, the increase was not statistically significant. Temporal analysis of tweet frequencies for the three themes of tweets showed that discussion increased after the onset of a fire but persisted typically no more than 2 weeks. For fires that burned for longer periods of more than a month, external events triggered climate discussions. Our findings contribute to identifying how the personal experience of wildfire shapes Twitter discussion related to climate change, and how these framings change over time during wildfire events, leading to insights into critical time points after wildfire for implementing message strategies to increase public engagement on climate change impacts and policy.

As social media is increasingly used by communities to understand and cope with environmental hazards, understanding how people use social media before, during, and after disasters can support disaster response and recovery efforts. This paper presents an empirical application of Houston et al.’s (Disasters 39:1–22, 2015) functional framework for disaster social media, using the case of Twitter use during and after Hurricane María. Our research aims to (1) identify the predominant patterns of Twitter usage and content dissemination during the Hurricane María crisis and (2) validate and refine the functional framework for disaster social media with a case study of the hurricane that hit Puerto Rico in 2017. We find that people in the US used Twitter mainly to access news of the hurricane, express emotions (both negative and positive), and to understand socio-political events shaping the response and recovery. Most tweets came from individuals rather than organizations, and most were sent as the hurricane was designated as category four and approached Puerto Rico, with far fewer posts after landfall. These findings highlight the importance of individuals sharing and accessing vital information when official outlets are absent or limited and the relatively short-lived attention to slow recovery processes.

Terrorism causes emotional reactions among public audiences, with downstream consequences for their well-being, attitudes and policy preferences. We utilise a novel approach which harnesses a unique dataset of Twitter activity from 324K users to precisely capture emotional responses to terrorism. Our results demonstrate that terrorist attacks induce dramatic spikes in various discrete emotions of a negative valence, which vary based on the characteristics of the attacks. Furthermore, we show that the terrorism-induced effects on emotions are accompanied by changes in feelings about immigration. [ABSTRACT FROM AUTHOR]

We study the prevalence and characteristics of toxic speech on Twitter in the run-up to the 2022 Italian General election. We analyzed over 8.5 M tweets shared by 450k unique users and employed a machine learning classifier to estimate the toxicity score of their messages. We found that supporters of different political coalitions exhibit varying levels of toxic speech, sending between 6-8% of toxic messages overall. Notably, Centre-Left politicians received more toxic messages on average, with the largest target of hate receiving over 15,000 toxic replies. We employed Generalized Linear Models to study factors that drive hate speech to political targets, finding that, importantly, politicians employing more abusive and harmful language are also more likely to attract more inflammatory speech. Our findings underscore the critical need for targeted interventions to address hate speech online, fostering healthier dialogue and safeguarding democratic discourse. [ABSTRACT FROM AUTHOR]

This study harnesses the linguistic diversity of Arabic dialects to create two expansive corpora from X (formerly Twitter). The Gulf Arabic Corpus (GAC-6) includes around 1.7 million tweets from six Gulf countries—Saudi Arabia, UAE, Qatar, Oman, Kuwait, and Bahrain—capturing a wide range of linguistic variations. The Saudi Dialect Corpus (SDC-5) comprises 790,000 tweets, offering in-depth insights into five major regional dialects of Saudi Arabia: Hijazi, Najdi, Southern, Northern, and Eastern, reflecting the complex linguistic landscape of the region. Both corpora are thoroughly annotated with dialect-specific seed words and geolocation data, achieving high levels of accuracy, as indicated by Cohen's Kappa scores of 0.78 for GAC-6 and 0.90 for SDC-5. The annotation process leverages AI-driven techniques, including machine learning algorithms for automated dialect recognition and feature extraction, to enhance the granularity and precision of the data. These resources significantly contribute to the field of Arabic dialectology and facilitate the development of AI algorithms for linguistic data analysis, enhancing AI system design and efficiency. The data provided by this research are crucial for advancing AI methodologies, supporting diverse applications in the realm of next-generation AI technologies. [ABSTRACT FROM AUTHOR]

Electoral cycles are highly charged, politically intense moments that influence public discourse. Political elites, citizens, and the traditional news media seek to generate opinions and interactions on social networks. This research is motivated by the following questions: What are the communicative and deliberative practices used in these spaces? Is it possible to identify the characteristics that --through deliberative conversations-- potentially foster or undermine democratic debate, particularly when using populist and polarizing discourse? Using a mixed methods approach, we apply a social network analysis tool to track conversations and identify the volume of political discourse (N=346,000). Using selective and staged filtering, we identified posts from nine Peruvian Twitter accounts during the September 2022 electoral campaign in Lima: three candidates, three media outlets, and three accounts with high levels of engagement. Our data comprised the comments from these nine accounts and was extracted using an Application Programming Interface (API). Subsequently, we carried out ethnographic content analysis on publications with more than 30 comments. In this phase, we analyzed Twitter comments using a codebook to identify deliberative practices and user responses. Our findings underscore the significant role of principal media outlets in shaping political conversation on Twitter. We also discovered that attempts to interact and deliberate were often overshadowed by heated comments attempting to impose opinions on others. Most importantly, our research reveals a widespread disillusionment with politics, political institutions, and even the political preferences of fellow citizens in Peru, highlighting a key challenge for democratic discourse on social media. [ABSTRACT FROM AUTHOR]

Recent advancements in cryptography and digital assets have triggered a profound transformation, causing a reorganization of the financial sector. The growing popularity of cryptocurrencies as investment instruments has raised concerns regarding the influencer mentions on social media impact on their price dynamics. This study assesses the discourse surrounding major cryptocurrencies and tokens within the decentralized finance (DeFi) and masternode sectors, examining its influence on prices and trading volumes during the period from 2019 to 2021. The study verifies the alignment between social media communications concerning these assets and their intended objectives. The analysis utilizes KHCoder software to explore topics with the highest public engagement, revealing that DeFi users frequently exhibit herd-like behavior by actively pursuing high-return farming trends, often at the expense of unique project attributes. In contrast, discourse surrounding masternode tokens more closely adheres to project objectives. The study employs panel models and Granger causality tests to investigate the relationship between cryptocurrency market performance, interactions on social media, and tweet sentiment. The findings substantiate a causal and temporal connection between micro-discourses on Twitter and asset prices, as well as trading volumes. [ABSTRACT FROM AUTHOR]

Copyright of Polish Statistician / Wiadomości Statystyczne is the property of State Treasury - Statistics Poland and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In the era of big data, effectively processing and understanding the vast quantities of brief texts on social media platforms like Twitter (X) is a significant challenge. This paper introduces a novel approach to automatic text summarization aimed at improving accuracy while minimizing redundancy. The proposed method involves a two-step process: first, feature extraction using term frequency–inverse document frequency (TF–IDF), and second, summary extraction through genetic optimized fully connected convolutional neural networks (GO-FC-CNNs). The approach was evaluated on datasets from the Kaggle collection, focusing on topics like FIFA, farmer demonstrations, and COVID-19, demonstrating its versatility across different domains. Preprocessing steps such as tokenization, stemming, stop word s removal, and keyword identification were employed to handle unprocessed data. The integration of genetic optimization into the neural network significantly improved performance compared to traditional methods. Evaluation using the ROUGE criteria showed that the proposed method achieved higher accuracy (98.00%), precision (98.30%), recall (98.72%), and F1-score (98.61%) than existing approaches. These findings suggest that this method can help create a reliable and effective system for large-scale social media data processing, enhancing data dissemination and decision-making. [ABSTRACT FROM AUTHOR]

In the rapidly evolving field of artificial intelligence, the importance of multimodal sentiment analysis has never been more evident, especially amid the ongoing COVID-19 pandemic. Our research addresses the critical need to understand public sentiment across various dimensions of this crisis by integrating data from multiple modalities, such as text, images, audio, and videos sourced from platforms like Twitter. Conventional methods, which primarily focus on text analysis, often fall short in capturing the nuanced intricacies of emotional states, necessitating a more comprehensive approach. To tackle this challenge, our proposed framework introduces a novel hybrid model, IChOA-CNN-LSTM, which leverages Convolutional Neural Networks (CNNs) for precise image feature extraction, Long Short-Term Memory (LSTM) networks for sequential data analysis, and an Improved Chimp Optimization Algorithm for effective feature fusion. Remarkably, our model achieves an impressive accuracy rate of 97.8%, outperforming existing approaches in the field. Additionally, by integrating the GeoCoV19 dataset, we facilitate a comprehensive analysis that spans linguistic and geographical boundaries, enriching our understanding of global pandemic discourse and providing critical insights for informed decision-making in public health crises. Through this holistic approach and innovative techniques, our research significantly advances multimodal sentiment analysis, offering a robust framework for deciphering the complex interplay of emotions during unprecedented global challenges like the COVID-19 pandemic. [ABSTRACT FROM AUTHOR]

The Paris Agreement represents the notable characteristics of global environmental governance, namely multi-actor and multi-level governance. Despite these changes, limited empirical studies have investigated the actors in global environmental governance who participate in political discourse on climate change. To comprehend whether or not the Paris Agreement's actors represent the transformation of global environmental governance, we adopted a big-data approach to analyze Twitter's climate-change discourse. The classification of actors provided an understanding of the diversity of the actors involved in providing public information. We also calculated the closeness centrality and identified the changes in the actors' influence, comparing the changes in the 5 years since the Agreement was adopted. The results showed an increased diversity of actor types. Non-state actors are not homogeneous in their changes, and Intergovernmental Organizations exhibited an increased influence. The findings indicated that the actors participating in the climate-change discourse post-Paris Agreement partially reflect the changes in global environmental governance. [ABSTRACT FROM AUTHOR]

Although the 2022 cryptocurrency market crash prompted despair among investors, the rallying cry, "wagmi" (We're all gonna make it.) emerged among cryptocurrency enthusiasts in the aftermath. Did cryptocurrency enthusiasts respond to this crash differently compared to traditional investors? Using natural language processing techniques applied to Twitter data, this study employed a difference-in-differences method to determine whether the cryptocurrency market crash had a differential effect on investor sentiment toward cryptocurrency enthusiasts relative to more traditional investors. The results indicate that the crash affected investor sentiment among cryptocurrency enthusiastic investors differently from traditional investors. In particular, cryptocurrency enthusiasts' tweets became more neutral and, surprisingly, less negative. This result appears to be primarily driven by a deliberate, collectivist effort to promote positivity within the cryptocurrency community ("wagmi"). Considering the more nuanced emotional content of tweets, it appears that cryptocurrency enthusiasts expressed less joy and surprise in the aftermath of the cryptocurrency crash than traditional investors. Moreover, cryptocurrency enthusiasts tweeted more frequently after the cryptocurrency crash, with a relative increase in tweet frequency of approximately one tweet per day. An analysis of the specific textual content of tweets provides evidence of herding behavior among cryptocurrency enthusiasts. [ABSTRACT FROM AUTHOR]

In recent decades, the field of statistical linguistics has made significant strides, which have been fueled by the availability of data. Leveraging Twitter data, this paper explores the English and Spanish languages, investigating their rank diversity across different scales: temporal intervals (ranging from 3 to 96 h), spatial radii (spanning 3 km to over 3000 km), and grammatical word ngrams (ranging from 1-grams to 5-grams). The analysis focuses on word ngrams, examining a time period of 1 year (2014) and eight different countries. Our findings highlight the relevance of all three scales with the most substantial changes observed at the grammatical level. Specifically, at the monogram level, rank diversity curves exhibit remarkable similarity across languages, countries, and temporal or spatial scales. However, as the grammatical scale expands, variations in rank diversity become more pronounced and influenced by temporal, spatial, linguistic, and national factors. Additionally, we investigate the statistical characteristics of Twitter-specific tokens, including emojis, hashtags, and user mentions, revealing a sigmoid pattern in their rank diversity function. These insights contribute to quantifying universal language statistics while also identifying potential sources of variation. [ABSTRACT FROM AUTHOR]

This study analyzes the distribution of content posted on Twitter in response to a specific event or crisis from 2014 to 2019. The aim of the study is to identify any shifts in the focus of the content and to explore the possible reasons for these changes. The findings suggest a shift from a disaster arena to a political arena over the six-year period. The initial years were dominated by content related to reporting on the situation, requesting help, and coordinating relief efforts, while the latter years saw an increase in content related to criticizing the government, appreciating government effort, and discussing social and political issues. The study provides insights into the changing nature of public responses to events and crises, and highlights the role of social media as a platform for political discussions. [ABSTRACT FROM AUTHOR]

Using daily‐level data on corporate social media activity, we show that investor attention generally increases when firms post on Twitter, Instagram and YouTube and that the effect is stronger during earnings announcement periods. We find that stock market liquidity improves when firms post on social media, but the effects are the most consistent for Twitter. Finally, we document that when firms miss earnings, they post more on social media if the magnitude of the bad news is small but remain silent when the magnitude is large. This strategic behaviour is prevalent across all three social media platforms. [ABSTRACT FROM AUTHOR]

Twitter is one of the well-known text-based social media that is often used to disseminate content. According to Katadata, Indonesia ranked fifth in the world in 2023. So many people or organizations want to make tweets go viral. Therefore, this research aims to develop a model that uses tweet data from the Indonesian language Twitter social media to categorize the level of virality. There are several tasks in classifying the level of virality, such as upsampling data, predicting sentiment and emotion, and text embedding. Upsampling data was carried out because the dataset used was an imbalanced dataset. Data upsampling, emotions, and text embedding is carried out using the bidirectional encoder representation from transformers (BERT) model. Meanwhile, sentiment prediction uses the Ro- bustly optimized BERT pretraining approach (RoBERTa). The results of text embedding, sentiment, emotion, will be combined with Twitter metadata then all features will be fed into the multi-layer perceptron (MLP) model to classifying the level of virality which is divided into 3 classes based on the number of retweets, namely low, medium and high. The proposed method produces an F1-score of 49% and an accuracy of 95% and performs better than the baseline model. [ABSTRACT FROM AUTHOR]

With the extensive technological advancements and expansion, the persistent issues regarding the creation and rapid dissemination of fake news have become a prevalent and recurrent concern. The manipulation of news content has critical repercussions, such as causing public mistrust, fear, harm, and misinformation. Addressing that, this study developed a supervised machine learning algorithm that can accurately classify social media data as fake news. The methodology of the proposed fake news detection model involved five main components: data acquisition from Twitter, data preprocessing, data transformation, model development using Naïve Bayes, decision tree, and support vector machine (SVM) and model evaluation using accuracy, precision, recall and F1-score. The results revealed that decision tree recorded the highest accuracy for both textual data (100%) and metadata (94.54%) and consistently outperformed both Naïve Bayes and SVM in terms of precision, recall, and F1-score metrics, with a score of 100% for the classification of textual data-based datasets. Regarding the metadata-based classification, decision tree also demonstrated excellent performance, with the highest F1-score of 94% for fake news data. Meanwhile, SVM exhibited the highest precision and recall performance for the metadata-based classification. Overall, the application of the decision tree classifier was deemed the most effective in Twitter fake news detection. [ABSTRACT FROM AUTHOR]

The aim of this study is to investigate how amplifiers are used in Twitter (now called " X ") during election campaigns. Specifically, the main purpose is to identify the role and engagement of Twitter amplifiers in the 2019 European elections, the visibility of political parties and leaders, and the way in which automated tools are used to manipulate public opinion by influencing voting decisions. The countries considered in the study are two economic powers of Western Europe, France and Germany, as well as two countries of the European South, which are affected by the economic and financial crisis, Greece and Italy. The countries from Southern Europe were included in the sample as they are often used by mass media as political campaign tools. This paper emphasizes the Twitter platform through which the data collection was implemented using the official API of the social networking tool, focusing on the 2019 European elections. We collected data on 88 party leaders and MEP candidates between 10 May and 30 May 2019, as well as on 44,651 accounts that retweeted them. We concluded using 237,813 election-related tweets and used network theory to analyze and visualize the data. The results demonstrate that all political parties use amplifiers to promote their tweets, and some use the same amplifiers between different countries. [ABSTRACT FROM AUTHOR]

Online social networks (OSNs) have reduced global boundaries, with Twitter enabling perspective sharing. Bot profile‐propagated false information misuse raises serious concerns. Considering this issue, we present our research on classifying Twitter accounts as "human" or "bot" using deep neural networks and transfer learning. Our proposed approach, TL‐PBot, stands for bot profile detection using transfer learning. The TL‐PBot framework utilizes Twitter account metadata such as follower count. Our TL‐PBot also incorporates text data from the Twitter description field as a feature. Word representation of the text data is achieved using Global Vectors (GloVe), a pre‐trained model. By employing user profile‐based features, we significantly reduce the overhead of feature engineering. The hybrid nature of the model enables it to effectively handle mixed‐type features, including text, binary, and numerical data. We design the network using long‐short‐term memory (LSTM) units. DNN model layers were trained, and the weights of the pre‐trained model layers were frozen to apply the transfer learning, resulting in reduced training time and improved bot profile detection accuracy. The performance of the proposed TL‐PBot is evaluated using publicly available datasets. The proposed approach is trained and tested on the same datasets and further evaluated on the validation datasets that were not used in the training phase, which is also a novelty in our approach. Comparative analysis with state‐of‐the‐art approaches demonstrates that the TL‐PBot approach achieves a higher accuracy of 98.07%, while excelling in precision of 99%, recall of 98%, f measure of 98.32%, and AUC of 0.99. Employing the transfer learning strategy resulted in an accelerated detection rate of 5.04 milliseconds, attesting to the effectiveness of this approach in enhancing computational efficiency. [ABSTRACT FROM AUTHOR]

The study empirically analyzed activism participants' roles drawn from the lens of social media affordance and identified the activism opinion leaders based on the framework of network connectivity, message diffusion, and semantic relevancy through the case of the #Marchforourlives Twitter network, which has been rebranded as X. The study defines the #Marchforourlives Twitter network as a co-created activism network in collaboration with different degrees of contributors, such as the core advocates, the advocates, the supporters, and the amplifiers. The results showed that a very small number of tweets created by the core advocates played significant roles due to their extensive adoption by other participants, while many other original tweets were never mentioned or retweeted in the network. This study disclosed the extensive proportion of amplifiers as 95.13% among the examined participants. The study findings suggest that creating core agenda tweets with high amplifiability might be critical for successful hashtag activism to attract like-minded masses as networked protesters. [ABSTRACT FROM AUTHOR]

The US federal menu labeling law, implemented on May 7 th 2018, required that restaurant chains post calorie counts on menu items. The purpose of this study was to analyze the change in public sentiment, using Twitter data, regarding eight restaurant chains before and after the calorie labeling laws implementation. Twitter data was mined from Twitters application programming interface (API) for this study from the calendar year 2018; 2016 and was collected as a control. We selected restaurant chains that had a range of compliance dates with the law. Tweets about each chain were filtered by brand-specific keywords, and Valence Aware Dictionary and sEntiment Reasoner (VADER) sentiment analysis was applied to receive a continuous compound score (-1-1) of how positive (1) or negative (-1) each tweet was. Controlled Interrupted Time Series (CITS) was performed with Ordinary Least Squares (OLS) Regression on 2018 and 2016 series of compound scores for each brand, and level and trend changes were calculated. Most restaurant chains that implemented the federal menu calorie labeling law experienced no change or a small change in level or trend in sentiment after they implemented labeling. Chains experienced mildly more negative sentiment right after the law was implemented, with attenuation of this effect over time. Calorie labeling did not have a strong effect on the publics perception of food brands over the long-term on Twitter and may imply the need for greater efforts to change the sentiment towards unhealthy restaurant chains.

Abstract Terrorism causes emotional reactions among public audiences, with downstream consequences for their well-being, attitudes and policy preferences. We utilise a novel approach which harnesses a unique dataset of Twitter activity from 324K users to precisely capture emotional responses to terrorism. Our results demonstrate that terrorist attacks induce dramatic spikes in various discrete emotions of a negative valence, which vary based on the characteristics of the attacks. Furthermore, we show that the terrorism-induced effects on emotions are accompanied by changes in feelings about immigration.

En este trabajo se presenta el estudio de los recursos de concesividad con función atenuante en actos argumentativos (van Eemeren, Grootendorst y Snoeck Henkemans, 2017) presentes en un corpus de tweets relacionados con la franja electoral chilena emitida para el plebiscito constitucional de octubre del 2020. La atenuación, entendida como un mecanismo retórico, destinado a convencer, persuadir y cuidar las relaciones sociales e interpersonales, o evitar su menoscabo, se relaciona con la noción de imagen de Goffman (1967). En este marco, una de las estrategias que posee función atenuante en actos argumentativos es la concesividad, esto es, una construcción que expresa una dificultad para la realización de lo enunciado en la oración principal, aunque esta dificultad no impediría su cumplimiento (RAE, 1973; Flamenco García, 1999; Nueva Gramática de Lengua Española, 2009). En este trabajo se abordan las construcciones concesivas en una muestra de 5 000 tuits producidos en el contexto del plebiscito por una nueva constitución en Chile en octubre del año 2020. Estas construcciones se exploran en dos principales dimensiones: la descripción de los movimientos concesivo-opositivos en actos argumentativos y el funcionamiento de estos movimientos en dichos actos a partir del empleo de partículas formales. El análisis sugiere que estas construcciones son altamente productivas —con diez tipos claramente diferenciados—, ya que los usuarios intentan autoproteger su imagen al reportar su opinión sobre temas polémicos, como lo es el debate político en relación con una nueva constitución en Chile.

Abstract We study the prevalence and characteristics of toxic speech on Twitter in the run-up to the 2022 Italian General election. We analyzed over 8.5 M tweets shared by 450k unique users and employed a machine learning classifier to estimate the toxicity score of their messages. We found that supporters of different political coalitions exhibit varying levels of toxic speech, sending between 6-8% of toxic messages overall. Notably, Centre-Left politicians received more toxic messages on average, with the largest target of hate receiving over 15,000 toxic replies. We employed Generalized Linear Models to study factors that drive hate speech to political targets, finding that, importantly, politicians employing more abusive and harmful language are also more likely to attract more inflammatory speech. Our findings underscore the critical need for targeted interventions to address hate speech online, fostering healthier dialogue and safeguarding democratic discourse.

Abstract In the rapidly evolving field of artificial intelligence, the importance of multimodal sentiment analysis has never been more evident, especially amid the ongoing COVID-19 pandemic. Our research addresses the critical need to understand public sentiment across various dimensions of this crisis by integrating data from multiple modalities, such as text, images, audio, and videos sourced from platforms like Twitter. Conventional methods, which primarily focus on text analysis, often fall short in capturing the nuanced intricacies of emotional states, necessitating a more comprehensive approach. To tackle this challenge, our proposed framework introduces a novel hybrid model, IChOA-CNN-LSTM, which leverages Convolutional Neural Networks (CNNs) for precise image feature extraction, Long Short-Term Memory (LSTM) networks for sequential data analysis, and an Improved Chimp Optimization Algorithm for effective feature fusion. Remarkably, our model achieves an impressive accuracy rate of 97.8%, outperforming existing approaches in the field. Additionally, by integrating the GeoCoV19 dataset, we facilitate a comprehensive analysis that spans linguistic and geographical boundaries, enriching our understanding of global pandemic discourse and providing critical insights for informed decision-making in public health crises. Through this holistic approach and innovative techniques, our research significantly advances multimodal sentiment analysis, offering a robust framework for deciphering the complex interplay of emotions during unprecedented global challenges like the COVID-19 pandemic.

Abstract The Paris Agreement represents the notable characteristics of global environmental governance, namely multi-actor and multi-level governance. Despite these changes, limited empirical studies have investigated the actors in global environmental governance who participate in political discourse on climate change. To comprehend whether or not the Paris Agreement’s actors represent the transformation of global environmental governance, we adopted a big-data approach to analyze Twitter’s climate-change discourse. The classification of actors provided an understanding of the diversity of the actors involved in providing public information. We also calculated the closeness centrality and identified the changes in the actors’ influence, comparing the changes in the 5 years since the Agreement was adopted. The results showed an increased diversity of actor types. Non-state actors are not homogeneous in their changes, and Intergovernmental Organizations exhibited an increased influence. The findings indicated that the actors participating in the climate-change discourse post-Paris Agreement partially reflect the changes in global environmental governance.

Abstract Twitter’s trending topics enable users to view the topics currently being discussed on the platform. Users can stay up to date with news, events, and conversations. However, the platform’s method of sorting tweets by time can make it hard to gather semantic information. To fully comprehend the various dimensions of trends and the diverse opinions surrounding them, users need to sift through a substantial number of results. Traditional techniques for content summarization, such as multi-document summarization, can facilitate information aggregation, categorization, and visualization of events, but there are two challenges. First, they fail to consider the topic’s polarity, which is essential to covering all aspects of the subject and incorporating less popular opinions. Second, some techniques only provide summaries at the topic level, potentially leaving out crucial dimensions that require representation in this summary. This research developed a novel summarization approach on Twitter which is known as ARAbic Trending SUMmarization (AraTSum). The proposed system generates the summary based on the extracted topics and aspects from the trend. The approach involves a topic sentiment-based technique that combines generative statistical Latent Dirichlet Allocation with a pre-trained model to automatically reflect the sentiments (negative or positive) of tweets in each topic; followed by extractive summarization algorithms in each cluster. The AraTSum was evaluated through several experiments on five different X datasets. The obtained results showed that AraTSum outperformed existing approaches on the ROUGE evaluation metric compared to state-of-the-art Twitter event summarizing algorithms. To ensure a comprehensive and accurate evaluation, three human experts were tasked with manually summarizing the utilized five datasets. The results demonstrated that the proposed AraTSum method is dependent on sentiment topical aspect analysis, and it enhances the summarization's performance.

With roots as a public health campaign in the United Kingdom, Dry January is a temporary alcohol abstinence initiative encouraging participants to abstain from alcohol use during the month of January. Dry January has become a cultural phenomenon, gaining increasing news media attention and social media engagement. Given the utility of capturing naturalistic discussions around health topics on social media, we examined Twitter chatter about Dry January and associated temporary abstinence experiences. Public tweets were collected containing the search terms dry january or dryjanuary posted between 15 December and 15 February across 3 years (2020-2). A random subsample stratified by year (n = 3145) was pulled for manual content analysis by trained coders. Final codebook accounted for user sentiment toward Dry January, user account type, and themes related to Dry January participation. Engagement metadata (e.g. likes) were also collected. Though user sentiment was mixed, most tweets expressed positive or neutral sentiment toward Dry January (74.7%). Common themes included encouragement and support for Dry January participation (14.1%), experimentation with and promotion of nonalcoholic drinks (14.0%), and benefits derived from Dry January participation (10.4%). While there is promise in the movement to promote positive alcohol-related behavior change, increased efforts to deliver the campaign within a public health context are needed. Health communication campaigns designed to inform participants about evidence-based treatment and recovery support services proven to help people quit or cut down on their drinking are likely to maximize benefits.

Abstract Objectives With the expansion of social networks such as Twitter, many experts share their opinions on various topics. The opinions of experts, who are also known as influencers, can be very influential. Combining these tweets and the historical prices of cryptocurrencies makes it possible to predict their price trends accurately. A Hybrid of RoBERTa deep neural network and BiGRU has been used for Sentiment Analysis (SA). Sentiments of tweets can be of great help to investors to understand the future behavior of the market and manage the stock portfolio. Unlike the tweets that are only extracted using the cryptocurrency name hashtag, the tweets of this dataset have specialized opinions and can determine the market trend. Data description The dataset created in this research concerns the opinions of more than 52 influencers (persons or companies) regarding eight cryptocurrencies. This dataset was collected through the Apify Twitter API for eight months, from February 2021 to June 2023. This dataset contains five Excel files and tweets, compound score, importance coefficient of each tweet, sentiment polarity, and historical prices of four cryptocurrencies: Bitcoin, Ethereum, Binance, and other information. These tweets cover the opinions of 52 influencers on more than 300 cryptocurrencies, although most comments are related to Bitcoin, Ethereum, and Binance. For this reason, three Excel files containing the historical prices of polarity and compound sentiment related to Bitcoin, Ethereum, and Binance cryptocurrencies have been placed separately in the dataset. The polarity of sentiment in these Excel shows the maximum number of polarities by applying the importance coefficient, which determines the dominant polarity of sentiment related to a particular day for the cryptocurrency.

Abstract The Ego Network Model (ENM) is a model for the structural organisation of relationships, rooted in evolutionary anthropology, that is found ubiquitously in social contexts. It takes the perspective of a single user (Ego) and organises their contacts (Alters) into a series of (typically 5) concentric circles of decreasing intimacy and increasing size. Alters are sorted based on their tie strength to the Ego, however, this is difficult to measure directly. Traditionally, the interaction frequency has been used as a proxy but this misses the qualitative aspects of connections, such as signs (i.e. polarity), which have been shown to provide extremely useful information. However, the sign of an online social relationship is usually an implicit piece of information, which needs to be estimated by interaction data from Online Social Networks (OSNs), making sign prediction in OSNs a research challenge in and of itself. This work aims to bring the ENM into the signed networks domain by investigating the interplay of signed connections with the ENM. This paper delivers 2 main contributions. Firstly, a new and data-efficient method of signing relationships between individuals using sentiment analysis and, secondly, we provide an in-depth look at the properties of Signed Ego Networks (SENs), using 9 Twitter datasets of various categories of users. We find that negative connections are generally over-represented in the active part of the Ego Networks, suggesting that Twitter greatly over-emphasises negative relationships with respect to “offline” social networks. Further, users who use social networks for professional reasons have an even greater share of negative connections. Despite this, we also found weak signs that less negative users tend to allocate more cognitive effort to individual relationships and thus have smaller ego networks on average. All in all, even though structurally ENMs are known to be similar in both offline and online social networks, our results indicate that relationships on Twitter tend to nurture more negativity than offline contexts.

This research seeks to understand the dynamics of controversy over wildlife preserving content on Twitter and how various actors voice their arguments on this issue. This research focuses on the process of argumentative discourse in the controversy over wildlife care content on Twitter. 353 tweets were found which produced 476 statements and 32 concepts summarized in four categories. The discourse that emerges is dominated by discourse voiced by actors who oppose the content of preserving wild animals. The first category is the reasons for not keeping and watching wildlife are consists of nine concepts. The second category, the reasons for maintaining and producing wildlife content, consists of seven concepts. The third category is the effects of keeping wildlife content consists of 3 concepts. Finally, the education category consists of six concepts. The results of this research show that netizens have the highest frequency of participation in voicing their opinions regarding the controversy over the content of wildlife conservation even without professional experience in the field of wildlife conservation. Apart from that, netizens have an important role in influencing the public's voice on this issue.

The study empirically analyzed activism participants’ roles drawn from the lens of social media affordance and identified the activism opinion leaders based on the framework of network connectivity, message diffusion, and semantic relevancy through the case of the #Marchforourlives Twitter network, which has been rebranded as X. The study defines the #Marchforourlives Twitter network as a co-created activism network in collaboration with different degrees of contributors, such as the core advocates, the advocates, the supporters, and the amplifiers. The results showed that a very small number of tweets created by the core advocates played significant roles due to their extensive adoption by other participants, while many other original tweets were never mentioned or retweeted in the network. This study disclosed the extensive proportion of amplifiers as 95.13% among the examined participants. The study findings suggest that creating core agenda tweets with high amplifiability might be critical for successful hashtag activism to attract like-minded masses as networked protesters.

The aim of this study is to investigate how amplifiers are used in Twitter (now called “X”) during election campaigns. Specifically, the main purpose is to identify the role and engagement of Twitter amplifiers in the 2019 European elections, the visibility of political parties and leaders, and the way in which automated tools are used to manipulate public opinion by influencing voting decisions. The countries considered in the study are two economic powers of Western Europe, France and Germany, as well as two countries of the European South, which are affected by the economic and financial crisis, Greece and Italy. The countries from Southern Europe were included in the sample as they are often used by mass media as political campaign tools. This paper emphasizes the Twitter platform through which the data collection was implemented using the official API of the social networking tool, focusing on the 2019 European elections. We collected data on 88 party leaders and MEP candidates between 10 May and 30 May 2019, as well as on 44,651 accounts that retweeted them. We concluded using 237,813 election-related tweets and used network theory to analyze and visualize the data. The results demonstrate that all political parties use amplifiers to promote their tweets, and some use the same amplifiers between different countries.

Elections are one of the key moments in a country's democracy. Indonesian elections have a significant impact on regional and global politics. Twitter being one of the popular social media platforms becomes a powerful tool for political campaigns. This makes it an ideal source to analyze public opinion during the 2024 general election, particularly the upcoming Presidential Election (Pilpres). IndoBERT is the model chosen to analyse the sentiment from the dataset in this study using a zero-shot learning approach. Based on the evaluation results, the accuracy value of the 2024 presidential election classification is 0.60 (60%), tends to predict with a good value in the positive label of 0.74 (74%) for F1-Score. This model is considered quite good at predicting negative labels but the results are not too optimal with a value of 0.49 (49%). Confusion Matrix in this IndoBERT model is more likely to label tweets with positive things, by detecting negative labels quite well.

Building upon earlier studies and contributions to knowledge, this research examines the lexemic formation processes and terminologies marked by hash tags. The goal is basically to demonstrate the modus operandi of formulating hash tags, as linguistic edifices, influencing people to act in a particular direction. To achieve this goal, the author consulted websites to harvest hash tag constructs at random in business, discrimination-cum-crime, culture, politics, and natural disaster spheres. The researcher further represents both hash tags and the formation recurrent patterns in statistics using tables and graphs. Besides the common constructs of imperative (#OccupyWallStreet as in # + Occupy + Wall + Street), declarative (#BlackLivesMatter as in # + Black + Life + s + Matter), and nominal (#eqnz as in # + e + q + n + z) elements, this study reveals that #LRNY depicts acronym in their morphological domains of # + L + R + N + Y. #McDStories explicates a word formation blending as in # + Mc + D + Stori + es to understand consumers’ perceptions. Very obvious are formations such as interro-declarative structures (#WhatPantherMeansToMe as in # + What + Black + Panther + Mean + s + To + Me,), verbo-nominal constructs (#AskObama as in # + Ask + Obama), and nominalised fabrications in a compounding pattern (#BringBackOurGirls as in # + Bring + Back + Our + Girl + s). Upon these fascinated liberty-constructions, the hashtag participants do not only have confidence to express their thoughts in the global terrain, users have the capacity to influence linguistic practices with novel formation ideas, illustrating things that people could do with words.

Abstract Aiming to debate the the power of the counter-mobilization that emerged on Twitter during the coronavirus pandemic in Brazil, we carried out a documental study on Twitter collecting texts and images associated with the #BolsonaroGenocida hashtag, between March 13 and June 6 of 2021, a period in which intense public demonstrations took place against the Bolsonaro’s government in the country. We analysed 1243 tweets and 503 images, identifying the main contents anchored and objectified by the hashtags, revealing the various dimensions that compose the social representations associated with Bolsonaro and his management during the pandemic. The tweets’ contents comprehend the social and political demonstration arising from part of the Brazilian people against the actions taken by the president of Brazil during the Covid-19 pandemic, which led to thousands of deaths that could have been avoided; regarding the images, the results point out an emotional anchoring of Bolsonaro’s image to death. It is therefore important to demarcate the nature of Twitter as an specific platform for the mobilization of SR through the sharing of images and texts, which ends up evoking emotions that contribute to accentuating the social identity of the group that is expressed.

Abstract Background There is sustained interest in understanding the perspectives of liver transplant recipients and living donors, with several qualitative studies shedding light on this emotionally charged subject. However, these studies have relied primarily on traditional semi-structured interviews, which, while valuable, come with inherent limitations. Consequently, there remains a gap in our comprehension of the broader public discourse surrounding living liver donation. This study aims to bridge this gap by delving into public conversations related to living liver donation through a qualitative analysis of Twitter (now X) posts, offering a fresh perspective on this critical issue. Methods To compile a comprehensive dataset, we extracted original tweets containing the hashtags “#donateliver” OR “#liverdonor”, all posted in English from January 1, 2012, to December 31, 2022. We then selected tweets from individual users whose Twitter (X) accounts featured authentic human names, ensuring the credibility of our data. Employing Braun and Clarke’s reflexive thematic analysis approach, the study investigators read and analysed the included tweets, identifying two main themes and six subthemes. The Health Policy Triangle framework was applied to understand the roles of different stakeholders involved in the discourse and suggest areas for policy improvement. Results A total of 361 unique tweets from individual users were analysed. The major theme that emerged was the persistent shortage of liver donors, underscoring the desperation faced by individuals in need of life-saving liver transplants and the urgency of addressing the organ shortage problem. The second theme delved into the experiences of liver donors post-surgery, shedding light on a variety of aspects related to the transplantation process, including the visibility of surgical scars, and the significance of returning to physical activity and exercise post-surgery. Conclusion The multifaceted experiences of individuals involved in the transplantation process, both recipients and donors, should be further studied in our efforts to improve the critical shortage of liver donors.

Abstract In this article, we present the findings of a comprehensive longitudinal social network analysis conducted on Twitter across four consecutive election campaigns in Spain, spanning from 2015 to 2019. Our focus is on the discernible trend of increasing partisan and ideological homogeneity within interpersonal exchanges on this social media platform, alongside high levels of networking efficiency measured through average retweeting. This diachronic study allows us to observe how dynamics of party competition might contribute to perpetuating and strengthening network ideological and partisan homophily, creating ‘epistemic bubbles’ in Twitter, yet showing a greater resistance to transforming them into ‘partisan echo-chambers.’ Specifically, our analysis reveals that the rise of a new radical right-wing party (RRP), Vox, has heightened ideological homogeneity among users across the entire ideological spectrum. However, this process has not been uniform. While users aligned with mainstream political parties consistently share content that reinforces in-party affinity, resulting in highly efficient ‘epistemic bubbles,’ the emergence of the RRP has given rise to a distinct group of users associated with the most extreme partisan positions, characterized by a notable proportion of out-partisan hostility content, which has fostered the creation of low-efficient 'partisan echo-chambers.'

In recent years, the opinion that the Eurovision Song Contest has become highly politicised is prevalent in the media and the popular voice, although not much research exists that can attest to this claim. In this work we conduct a case study that applies sentiment and discourse analysis methodologies to the assessment of political opinions in social media regarding this artistic and social event. The main objective is to explore to what extent and in what form this supposed politicisation has an expression on Twitter, as illustrated by the cases of artists Sam Ryder and James Newman, the United Kingdom’s representatives in the 2022 and 2021 editions of the contest, respectively. We examine references to two historical-political contexts that have had a severe impact on the European society over the last few years, and which have determined, among many other social aspects, the reception of Eurovision results ever since they took place: Brexit and the Russian invasion of Ukraine.

This article provides, first, a description then an analysis of the overall framework of a communication situation on X/Twitter. By opting for an analysis according to the model of settings of Brown and Fraser, we identify the contours of the notion of reception as a component of the more developed diagram of communication. Apart from the particularities marking computer-mediated communications, the analysis we carried out allows us to dissect the context and reveals how reception must be adapted to interaction on Twitter.

Media sosial Twitter adalah platform yang populer di Indonesia untuk berkomunikasi dan mendapatkan informasi dengan cepat. Hal ini memungkinkan masyarakat dengan mudah mengungkapkan opini dan sentimen mereka. Penelitian ini berfokus pada perbandingan algoritma TextBlob dan Naïve Bayes dalam menganalisis sentimen masyarakat. Temuan menunjukkan bahwa TextBlob mengklasifikasikan sebagian besar tweet sebagai positif, sementara Naïve Bayes menunjukkan kecenderungan yang serupa dengan akurasi sebesar 78,18%. Dari analisis TextBlob, sekitar 50,98% komentar menunjukkan sentimen positif, 16,01% negatif, dan 33,33% netral. Dengan menggunakan kedua algoritma ini, penelitian berhasil mengidentifikasi sentimen masyarakat dengan akurasi yang baik, menunjukkan distribusi yang jelas antara sentimen positif, netral, dan negatif.

Abstract Our social identities determine how we interact and engage with the world surrounding us. In online settings, individuals can make these identities explicit by including them in their public biography, possibly signaling a change in what is important to them and how they should be viewed. While there is evidence suggesting the impact of intentional identity disclosure in online social platforms, its actual effect on engagement activities at the user level has yet to be explored. Here, we perform the first large-scale study on Twitter that examines behavioral changes following identity disclosure on Twitter profiles. Combining social networks with methods from natural language processing and quasi-experimental analyses, we discover that after disclosing an identity on their profiles, users (1) tweet and retweet more in a way that aligns with their respective identities, and (2) connect more with users that disclose similar identities. We also examine whether disclosing the identity increases the chance of being targeted for offensive comments and find that in fact (3) the combined effect of disclosing identity via both tweets and profiles is associated with a reduced number of offensive replies from others. Our findings highlight that the decision to disclose one’s identity in online spaces can lead to substantial changes in how they express themselves or forge connections, with a lesser degree of negative consequences than anticipated.

ÖZ:Bu çalışmada amaç Covid-19 döneminde Twitter ile hisse senedi fiyatları arasındaki ilişkiyi incelemektir. Bunun için kısıtlamaların yaşandığı 01.10.2020 ile 01.07.2021 tarih aralığında BIST Yıldız Pazar’da yer alan ve imalat-sanayi sektöründen üç şirketin (Akçansa Çimento Sanayi ve Ticaret A.Ş., Coco-Cola İçecek A.Ş., Goodyear Lastikleri T.A.Ş.) hisse senedi günlük kapanış fiyatları ve Twitter üzerinden bu şirketlerle ilgili paylaşılan Türkçe tweetlerin günlük duygu durumu (negatif ve pozitif) dikkate alınmıştır. Tweetlerin duygu analizinde Turkish Bert modelinden yararlanılmıştır. Turkish Bert modeli bir metni negatif ve pozitif olarak duygulara ayırmada %95’in üzerinde bir başarıya sahiptir. Duygu analizi sonucunda şirketler ile ilgili elde edilen negatif ve pozitif tweet sayıları ile günlük hisse senedi kapanış fiyatları arasındaki ilişki Granger nedensellik testi ile analiz edilmiştir. Analizin sonucunda AKCNS hisse senedi getirisi ile negatif tweetler arasında çift yönlü bir ilişki dolayısıyla bu değişkenlerin birbirini etkilediği saptanmıştır. AKCNS hisse senedi getirisi pozitif tweetleri etkilediği fakat pozitif tweetlerden etkilenmediği gözlenmiştir. CCOLA ve GOODY hisse senedi getirilerinin ise sadece pozitif ve negatif tweetleri etkilediği tespit edilmiştir. Çalışma, daha önce yerli literatürde Covid-19 dönemi için Twitter ile hisse senedi fiyatları arasındaki ilişki incelenmemesi bakımından yatırımcılara ve işletmelere önemli bilgiler sunmaktadır. ABSTRACT:The purpose of this study is to examine the relationship between Twitter and stock prices during the Covid-19 period. Therefore, the daily closing prices of the stocks of three companies (Akçansa Cement Industry and Trade Inc., Coca-Cola Beverages Joint Stock Company, Goodyear Tires T. Joint Stock Company) from the manufacturing-industry sector, in the BIST Star Market between the dates of 01.10.2020 and 01.07.2021, when the restrictions were experienced, and the daily mood (negative and positive) of the Turkish tweets shared about these companies on Twitter was taken into account. Turkish Bert model was used in sentiment analysis of tweets. The Turkish Bert model has a success rate of over 95% in classifying texts into negative and positive emotions. The relationship between the number of negative and positive tweets about companies as a result of sentiment analysis and daily stock closing prices was analyzed with the Granger causality test. As a result of the analysis, it was determined that these variables affected each other due to a bidirectional relationship between AKCNS stock returns and negative tweets. It has been observed that AKCNS stock returns affected positive tweets but not vice versa. CCOLA and GOODY stock returns were found to influence only positive and negative tweets. The study provides important information to investors and businesses, as the relationship between Twitter and stock prices for the Covid-19 period has not been examined in the domestic literature before.

Abstract On February 28th, shortly after the Russian invasion of Ukraine on February 24th, Twitter announced the expansion of its labelling policy for “Russia state-affiliated media”, in order to address disinformation in favour of the Russian government.. While this ‘soft’ approach does not include the removal of content, it entails issues for freedom of expression and information. This article investigates the consequences of this labelling policy for the range and impact of accounts labelled “Russia state-affiliated media” during the Ukrainian war. Using an iterative detection method, a total of 90 accounts of both media outlets and individual journalists with this label were identified. The analysis of these accounts’ information and timeline, as well as the comparison of the impact of their tweets before and after February 28th with an ARIMA model, strongly suggests, that this policy, despite its limited scope, could have contributed to a reduction in the impact of the sampled tweets, among other concurrent events. These results provide empirical evidence to guide critical reflection on this content moderation policy.

Abstract Across Europe, negative public opinion has and may continue to limit the deployment of renewable energy infrastructure required for the transition to net‐zero energy systems. Understanding public sentiment and its spatio‐temporal variations is as such important for decision‐making and for developing socially accepted energy systems. In this study, we apply a sentiment classification model based on a machine learning framework for natural language processing, NorBERT, on data collected from Twitter between 2006 and 2022 to analyse the case of wind power opposition in Norway. From the 68,828 tweets with geospatial information, we show how discussions about wind power intensified in 2018/2019 together with a trend of more negative tweets up until 2020, both on a regional level and for Norway as a whole. Furthermore, we find weak geographical clustering in our data, indicating that discussions are country wide and not dominated by specific regional events or developments. Twitter data allow for detailed insight into the temporal nature of public sentiments and extending this research to additional case studies of technologies, countries and sources of data (e.g. newspapers, other social media) may prove important to complement traditional survey research and the understanding of public sentiment.

The Ego Network Model (ENM) is a model for the structural organisation of relationships, rooted in evolutionary anthropology, that is found ubiquitously in social contexts. It takes the perspective of a single user (Ego) and organises their contacts (Alters) into a series of (typically 5) concentric circles of decreasing intimacy and increasing size. Alters are sorted based on their tie strength to the Ego, however, this is difficult to measure directly. Traditionally, the interaction frequency has been used as a proxy but this misses the qualitative aspects of connections, such as signs (i.e. polarity), which have been shown to provide extremely useful information. However, the sign of an online social relationship is usually an implicit piece of information, which needs to be estimated by interaction data from Online Social Networks (OSNs), making sign prediction in OSNs a research challenge in and of itself. This work aims to bring the ENM into the signed networks domain by investigating the interplay of signed connections with the ENM. This paper delivers 2 main contributions. Firstly, a new and data-efficient method of signing relationships between individuals using sentiment analysis and, secondly, we provide an in-depth look at the properties of Signed Ego Networks (SENs), using 9 Twitter datasets of various categories of users. We find that negative connections are generally over-represented in the active part of the Ego Networks, suggesting that Twitter greatly over-emphasises negative relationships with respect to "offline" social networks. Further, users who use social networks for professional reasons have an even greater share of negative connections. Despite this, we also found weak signs that less negative users tend to allocate more cognitive effort to individual relationships and thus have smaller ego networks on average. All in all, even though structurally ENMs are known to be similar in both offline and online social networks, our results indicate that relationships on Twitter tend to nurture more negativity than offline contexts. [ABSTRACT FROM AUTHOR]

With this work, we hypothesize that semantically enriching a user's text corpus using backtranslation and expansion modules can improve performance for author profiling tasks. To perform this textual enrichment, we translate an author's representative text. Translations are made from one language—the source language—into another—the target language—and then back to the original one. Finally, we expand an author's text by integrating the original version with the back‐translated one. Our framework includes these backtranslation and expansion modules followed by a SOTA classifier successfully employed for text classification. The framework is tested on three author profiling datasets from the last three years' shared tasks on fake news, hate speech, irony and stereotypes detection hosted at the CLEF conference for the PAN Lab. This work is an extension of our previous one where we just presented our main idea. Here we improve our framework, and we also investigate more languages and more datasets. Finally, a qualitative analysis is provided. The results confirm that the backtranslation and expansion add‐on modules improve model performance on all three datasets evaluated. [ABSTRACT FROM AUTHOR]