Online Social Networks and Media最新文献

This article explores how political actors use the emotions of fear and anger in what we call fear-anger contests. Our theory distinguishes between governmental and populist actors and posits that, in a contest for media attention and the hearts and minds of citizens, populists pursue a politics of anger whereas governmental actors pursue a politics of fear. To evaluate the theory, we examine two episodes of contentious politics: the 2016 Brexit referendum and the election of Donald Trump in the same year. We rely on automated sentiment analysis, using machine learning and emotion dictionaries to examine a dataset of social media posts on Twitter. In the case of Brexit, we find a fear-anger contest between Remain (“Project Fear”) and Leave (“Project Anger”). In the case of the 2016 US presidential election, we find a negativity contest where both parties reinforce each other's negative emotions.

News media has long been an ecosystem of creation, reproduction, and critique, where news outlets report on current events and add commentary to ongoing stories. Understanding the dynamics of news information creation and dispersion is important to accurately ascribe credit to influential work and understand how societal narratives develop. These dynamics can be modelled through a combination of information-theoretic natural language processing and networks; and can be parameterised using large quantities of textual data. However, it is challenging to see “the wood for the trees”, i.e., to detect small but important flows of information in a sea of noise. Here we develop new comparative techniques to estimate temporal information flow between pairs of text producers. Using both simulated and real text data we compare the reliability and sensitivity of methods for estimating textual information flow, showing that a metric that normalises by local neighbourhood structure provides a robust estimate of information flow in large networks. We apply this metric to a large corpus of news organisations on Twitter and demonstrate its usefulness in identifying influence within an information ecosystem, finding that average information contribution to the network is not correlated with the number of followers or the number of tweets. This suggests that small local organisations and right-wing organisations which have lower average follower counts still contribute significant information to the ecosystem. Further, the methods are applied to smaller full-text datasets of specific news events across news sites and Russian troll accounts on Twitter. The information flow estimation reveals and quantifies features of how these events develop and the role of groups of trolls in setting disinformation narratives. In summary, this work provides a new methodology for examining the information transmitted between content producers in any connected system of natural language, a toolkit with applications to the many networked discourses of our online world.

The continuous proliferation of social media platforms and the exponential increase in users’ engagement are impacting social behavior and leading to various challenges, including the detection and identification of key influencers. In fact the opinions of these influencers are at the core of decision-making strategies, and are leading trends on the virtual social media landscape. Moreover, influencers might play a crucial role when it comes to misinformation and conspiracy during sensitive, controversial and trending events. However, due to the dynamic and unrestricted nature of social media, and diversity of targeted topics and audiences, identifying and ranking key influencers that are impactful, credible, and knowledgeable about their specialist topic or event remains an evolving and open research paradigm. In this paper, we address the aforementioned problem by proposing a novel influence rating and ranking scheme to identify key and highly influential users for a certain event over Twitter using a mixed theme/event based approach while considering historical data and profile reputation. We further apply our approach to a global pandemic case study, the novel Coronavirus, and conduct performance analysis. The presented experimental results and theoretical analysis explore the relevance of our proposed scheme for identifying and ranking reputable and theme/event related influencers.

Fake accounts are the primary means for misuse and abuse of social media platforms, giving rise to coordinated inauthentic behaviors. Despite ongoing efforts to limit their exploitation, ready-to-use fake accounts can be found for sale on several underground markets. For the present study, we devised an innovative approach to detect accounts for sale on an underground market. Between June 2019 and July 2021, we detected more than 60,000 fake accounts, which we continuously tracked for changes in profile information and timeline. Afterward, we focused on the 23,579 accounts that produced at least one tweet in 2020, identifying the main characteristics like the most used names and profile descriptions. Also, we analyzed more than five million interactions, including mentions, replies, retweets, and the use of hashtags and URLs in tweets. These analyses exposed behavioral patterns indicating coordination, like using similar profile names or retweeting the same user. In particular, we spotted four coordinated campaigns, whose behavior ranged from attempting to influence the political debate in Buenos Aires to aggressive spam activity aimed at scamming cryptocurrency users or advertising counterfeit goods.

The most developing issue in analysing complex networks and graph mining is link prediction, which can be studied for both content and structural-based analysis in a social network. Link prediction deals with the prediction of missing links by determining whether a link can be created between two nodes in a future snapshot of a given directed weighted graph. Existing link prediction methods are only studied for unsigned graphs and work on principles of the common neighbourhood. However, the link prediction problem can also be studied for signed graphs where signed links can give an interesting insight into user associations. Obstruction of studies in this domain is caused by imbalance of class, i.e., positive links are frequent than negative ones, and forbearance of hidden communities. A signed network is a combination of dense and hidden communities. A hidden community structure is overlooked by majority of existing applications, taking dense community structure, i.e., one whole graph as input for developing a link prediction model. Hence, complete network information is required by majority of existing approaches, which seems unrealistic in modern social network analytics. In this article, we exploit hidden network communities to address link prediction problem in the signed network, focusing on negative links. A number of observation were made regarding negative links and a principle ensemble framework, i.e., $E$ NeLp, is proposed, having two phases, i.e, network embedding and classifier prediction. Using a probabilistic embedding framework, network representation of hidden signed communities is learned, which were then passed to a learning classifier to predict negative links, keeping intact the ensemble framework. Despite the limited availability of signed network datasets, an extensive experimental study was performed to evaluate $E$ NeLp pertinency, robustness, and scalability. The performance result shows that $E$ NeLp can be a promising consideration for addressing link prediction tasks in signed networks and gives encouraging results.

Several innovative applications could be advantageously placed within social networks, to be effective, attractive, and pervasive. Examples of application domains that could benefit from social networks are e-democracy, e-participation, online surveys, crowdsourcing, and proximity-based services. In all the above cases, users’ anonymity could represent a considerable added value or could be even necessary to develop the service. We observe that all the above domains are characterized by the fact that only a few asynchronous messages should be exchanged. Therefore, we do not need the full communication power of anonymous communication networks, in which low-latency and connection-oriented communication should be supported. On the other hand, unlike communication networks, the threat model we have to consider assumes the presence of an adversary (represented by an honest-but-curious social network provider) able to monitor the entire flow of the exchanged messages. In this paper, we propose an anonymous communication protocol for short communications in social networks, based on a collaborative approach. The proposed solution hides from the social network provider not only the content of the messages but also the communication itself, which, per se, can result in considerable privacy leakage (think of the case of proximity testing performed between two users). This enables the implementation, within the social network, of the above-mentioned applications. To give a concrete proof of this statement, we develop a privacy-preserving proximity-based solution which provides both symmetric and asymmetric proximity testing entirely within social networks.

Depression is a challenge to public health, frequently related to disability and one of the reasons that lead to suicide. Many of the ones who suffer depression use social media to obtain information or even to talk about their problem. Some studies have proposed to detect potentially depressive users in these online environments. However, unsatisfactory effectiveness is still a barrier to practical application. Hence, we propose a method of early detection of depression in social media based on a convolutional neural network in combination with context-independent word embeddings and Early and Late Fusion approaches. These approaches are experimentally evaluated, considering the importance of the underlying emotions encoded in the emoticons. The results show that the proposed method was able to detect potentially depressive users, reaching a precision of 0.76 with equivalent or superior effectiveness in relation to many baselines ($F_1(0.71)$). In addition, the semantic mapping of emoticons allowed to obtain significantly better results, including higher recall and precision with gains of 46.3% and 32.1%, respectively. Regarding the baseline word embedding approach, the higher recall and precision gains of our semantic mapping of emoticons were 14.5% and 40.8%. In terms of overall effectiveness, this work advanced the state-of-the-art, considering both individual embeddings and the fusion-based approaches. Moreover, it is demonstrated that emotions expressed by depressed people and encoded through emoticons are important suggestive evidence of the problem and a valuable asset for early detection.

With the advent of blockchain technology, new Online Social Networks (OSNs) were proposed under the name of Blockchain Online Social Media (BOSM). Among the most well-known BOSMs, the introduction of blockchain is geared towards providing a decentralised social platform, together with an auditable and transparent rewarding system. While some current BOSMs have gathered a large set of dedicated users, they can hardly fight against the hegemony of the most well known centralised platforms, such as Twitter or Facebook. Yup tries to overcome this issue by proposing a rewarding system that can integrate with existing platforms. Its rewarding system keeps track of any piece of content uniquely identified by an URL, so its usage is not restricted to OSNs only. Given its unique approach, its rewarding system represents an important case study that, to the best of our knowledge, is not covered in the literature. In this paper, we close this gap by presenting the rewarding system provided by Yup and understanding its implications on the social activity of the users. Our analyses uncover that the voting activity favours the creation of echo chambers, and that the rewarding system is unfair. Additionally, we identify some limitations that can help design new rewarding systems.

Twitter as a platform is used for news dissemination, with high volumes of campaigning and populism. This situation coincides with the growth of audiences who embrace social media as their primary news source. In general, effects like the deterioration of political education, misinformation, or ideological segregation then arguably represent a tremendous risk for democratic societies.

We analyze a comprehensive data set of the German-speaking Twitter community – a concise, well-defined Twitter population – to understand the extent and form of consumption of controversial news.

Our results affirm a high interest of German Twitter users in daily news and corresponding discussions. In-depth studies on the behavior, including tweeting- and grouping patterns, revealed the emergence of a new, more self-assured form of echo chambers.

Community detection is one of the primary problems in social network analysis and this problem has more challenges in attributed social networks. The purpose of community detection in attributed social networks is to discover communities with not only homogeneous node properties but also adherent structures. Although community detection has been extensively studied, attributed community detection of large social networks with a large number of attributes remains a vital challenge. To address this challenge, in this paper a novel attributed community detection method is developed by integration of feature weighting with node centrality techniques. The developed method includes two main phases: (1) Weight Matrix Calculation, (2) Label Propagation Algorithm-based Attributed Community Detection. The aim of the first phase is to calculate the weight between two linked nodes using structural and attribute similarities, while, in the second phase, an improved label propagation algorithm-based community detection method in the attributed social network is proposed. The purpose of the second phase is to detect different communities by employing the calculated weight matrix and node popularity. After implementing the proposed method, its performance is compared with several other state of the art methods using some benchmarked real-world datasets. The results indicate that the developed method outperforms several other state-of-the-art methods and ascertain the effectiveness of the developed method for attributed community detection.