Search results

Twitter fake accounts refer to bot accounts created by third-party organizations to influence public opinion, commercial propaganda or impersonate others. The effective identification of bot accounts is conducive to accurately judge the disseminated information for the public. However, in actual fake account identification, it is expensive and inefficient to manually label Twitter accounts, and the labeled data are usually unbalanced in classes. To this end, the authors propose a novel framework to solve these problems.

In the proposed framework, the authors introduce the concept of semi-supervised self-training learning and apply it to the real Twitter account data set from Kaggle. Specifically, the authors first train the classifier in the initial small amount of labeled account data, then use the trained classifier to automatically label large-scale unlabeled account data. Next, iteratively select high confidence instances from unlabeled data to expand the labeled data. Finally, an expanded Twitter account training set is obtained. It is worth mentioning that the resampling technique is integrated into the self-training process, and the data class is balanced at the initial stage of the self-training iteration.

The proposed framework effectively improves labeling efficiency and reduces the influence of class imbalance. It shows excellent identification results on 6 different base classifiers, especially for the initial small-scale labeled Twitter accounts.

This paper provides novel insights in identifying Twitter fake accounts. First, the authors take the lead in introducing a self-training method to automatically label Twitter accounts from the semi-supervised background. Second, the resampling technique is integrated into the self-training process to effectively reduce the influence of class imbalance on the identification effect.

Social bots are prevalent on social media. Malicious bots can severely distort the true voices of customers. This paper aims to examine social bots in the context of big data of user-generated content. In particular, the author investigates the scope of information distortion for 24 brands across seven industries. Furthermore, the author studies the mechanisms that make social bots viral. Last, approaches to detecting and preventing malicious bots are recommended.

A Twitter data set of 29 million tweets was collected. Latent Dirichlet allocation and word cloud were used to visualize unstructured big data of textual content. Sentiment analysis was used to automatically classify 29 million tweets. A fixed-effects model was run on the final panel data.

The findings demonstrate that social bots significantly distort brand-related information across all industries and among all brands under study. Moreover, Twitter social bots are significantly more effective at spreading word of mouth. In addition, social bots use volumes and emotions as major effective mechanisms to influence and manipulate the spread of information about brands. Finally, the bot detection approaches are effective at identifying bots.

As brand companies use social networks to monitor brand reputation and engage customers, it is critical for them to distinguish true consumer opinions from fake ones which are artificially created by social bots.

This is the first big data examination of social bots in the context of brand-related user-generated content.

This paper aims to examine how banks in Tanzania have been vulnerable to money laundering activities and how the banking institutions have been implicated in enabling or aiding the commission of money laundering offences, and highlights the banks’ failure or inability to prevent, detect and thwart money laundering committed through their financial systems.

The paper explores Tanzania’s anti-money laundering law and analyzes non-law factors that make the banks exposed to money laundering activities. It looks at law-related, political and economic circumstances that impinge on the banks’ efficacy to tackle money laundering offences committed through their systems. The data are sourced from policy documents, statutes, case law and literature from Tanzania and other jurisdictions.

Both law-related and non-law factors create an enabling environment for the commission of money laundering offences, and this exposes banks in Tanzania to money laundering activities. Some banks have been implicated in enabling or aiding money laundering offences. These banks have abdicated their obligations to fight against money laundering. This is attributed to the fact that the banks’ internal anti-money laundering policies, regulations and procedures are inefficient, and Tanzania’s legal framework is generally ineffective to tackle money laundering offences.

This paper uncovers a multi-faceted nature of money laundering affecting banks in Tanzania. It is recommended that Tanzania’s anti-money laundering policy should address law-related, political, economic and other factors that create an enabling environment for the commission of money laundering offences. Tanzania’s anti-money laundering law should be reformed to enhance its efficacy and, lastly, banks should reinforce their internal anti-money laundering policies and regulations and policies.

The purpose of this paper is to examine one of the largest data sets on the hashtag use of #fakenews that comprises over 14m tweets sent by more than 2.4m users.

Tweets referencing the hashtag (#fakenews) were collected for a period of over one year from January 3 to May 7 of 2018. Bot detection tools were employed, and the most retweeted posts, most mentions and most hashtags as well as the top 50 most active users in terms of the frequency of their tweets were analyzed.

The majority of the top 50 Twitter users are more likely to be automated bots, while certain users’ posts like that are sent by President Donald Trump dominate the most retweeted posts that always associate mainstream media with fake news. The most used words and hashtags show that major news organizations are frequently referenced with a focus on CNN that is often mentioned in negative ways.

The research study is limited to the examination of Twitter data, while ethnographic methods like interviews or surveys are further needed to complement these findings. Though the data reported here do not prove direct effects, the implications of the research provide a vital framework for assessing and diagnosing the networked spammers and main actors that have been pivotal in shaping discourses around fake news on social media. These discourses, which are sometimes assisted by bots, can create a potential influence on audiences and their trust in mainstream media and understanding of what fake news is.

This paper offers results on one of the first empirical research studies on the propagation of fake news discourse on social media by shedding light on the most active Twitter users who discuss and mention the term “#fakenews” in connection to other news organizations, parties and related figures.

In order to further advance the research of social bots, based on the latest research trends and in line with international research frontiers, it is necessary to understand the global research situation in social bots.

Choosing Web of Science™ Core Collections as the data sources for searching social bots research literature, this paper visually analyzes the processed items and explores the overall research progress and trends of social bots from multiple perspectives of the characteristics of publication output, major academic communities and active research topics of social bots by the method of bibliometrics.

The findings offer insights into research trends pertaining to social bots and some of the gaps are also identified. It is recommended to further expand the research objects of social bots in the future, not only focus on Twitter platform and strengthen the research of social bot real-time detection methods and the discussion of the legal and ethical issues of social bots.

Most of the existing reviews are all for the detection methods and techniques of social bots. Unlike the above reviews, this study is a systematic literature review, through the method of quantitative analysis, comprehensively sort out the research output in social bots and shows the latest research trends in this area and suggests some research indirections that need to be focused in the future. The findings will provide references for subsequent scholars to research on social bots.

The peer review history for this article is available at: https://publons.com/publon/10.1108/OIR-06-2021-0336.

The purpose of this study is to examine how factual information and misinformation are being shared on Twitter by identifying types of social media users who initiate the information diffusion process.

This study used a mixed methodology approach to analyze tweets with COVID-19-related hashtags. First, a social network analysis was conducted to identify social media users who initiate the information diffusion process, followed by a quantitative content analysis of tweets by users with more than 5K retweets to identify what COVID-19 claims, factual information, misinformation and disinformation was shared on Twitter.

Results found very little misinformation and disinformation distributed widely. While health experts and journalists shared factual COVID-19-related information, they were not receiving optimum engagement. Tweets by citizens focusing on personal experience or opinions received more retweets and likes compared to any other sender type. Similarly, celebrities received more replies than any other sender type.

This study helps medical experts and government agencies understand the type of COVID-19 content and communication being shared on social media for population health purposes.

This study offers insight into how social media users engage with COVID-19-related information on Twitter and offers a typology of categories of information shared about the pandemic.

The peer review history for this article is available at: https://publons.com/publon/10.1108/OIR-03-2021-0143/.

The purpose of this paper is to understand why some US Senators have more low-quality followers than others and the potential impact of low-quality followers on understanding constituent preferences.

For each US Senator, data on Twitter followers was matched with demographic characteristics proven to influence behavior. An OLS regression model evaluated why some Senators attract more low-quality followers than others. Then, observations on the impact of low-quality followers were discussed along with potential effects on information gathering and constituent representation.

This study finds that total followers, ideology and length of time on Twitter are all significant predictors of whether a Senator might attract low-quality followers. Low-quality followers can have wide-ranging implications on Senator’s use of social media data to represent constituents and develop public policy.

The data set only includes Senators from the 115th Congress (2017–2018). As such, future research could expand the data to include additional Senators or members of the House of Representatives.

Information is essential in any decision-making environment, including legislatures. Understanding why some users, particularly public opinion leaders, attract more low-quality social media followers could help decision-makers better understand where information is coming from and how they might choose to evaluates its content.

This study finds two practical implications for public opinion leaders, including Senators. First, accounts must be actively monitored to identify and weed-out low-quality followers. Second, users need to be wary of disinformation and misinformation and they need to develop strategies to identify and eliminate it from the collection of follower preferences.

This study uses a unique data set to understand why some Senators have more low-quality followers than others and the impact on information gathering. Other previous studies have not addressed this issue in the context of governmental decision-making or constituent representation.

By analyzing journal articles with high citation counts but low Twitter mentions and vice versa, the purpose of this paper is to provide an overall picture of differences between citation counts and Twitter mentions of academic articles.

Citation counts from the Web of Science and Twitter mentions of psychological articles under the Social Science Citation Index collection were collected for data analysis. An approach combining both statistical and simple content analysis was adopted to examine important factors contributing to citation counts and Twitter mentions, as well as the patterns of tweets mentioning academic articles.

Compared to citation counts, Twitter mentions have stronger affiliations with readability and accessibility of academic papers. Readability here was defined as the content size of articles and the usage of jargon and scientific expressions. In addition, Twitter activities, such as the use of hashtags and user mentions, could better facilitate the sharing of articles. Even though discussions of articles or related social phenomena were spotted in the contents of tweets, simple counts of Twitter mentions may not be reliable enough for research evaluations due to issues such as Twitter bots and a deficient understanding of Twitter users’ motivations for mentioning academic articles on Twitter.

This study has elaborated on the differences between Twitter mentions and citation counts by comparing the characteristics of Twitter-inclined and citation-inclined articles. It provides useful information for interested parties who would like to adopt social web metrics such as Twitter mentions as traces of broader engagement with academic literature and potential suggestions to increase the reliability of Twitter metrics. In addition, it gives specific tips for researchers to increase research visibility and get attention from the general public on Twitter.

The proliferation of bots in social networks has profoundly affected the interactions of legitimate users. Detecting and rejecting these unwelcome bots has become part of the collective Internet agenda. Unfortunately, as bot creators use more sophisticated approaches to avoid being discovered, it has become increasingly difficult to distinguish social bots from legitimate users. Therefore, this paper proposes a novel social bot detection mechanism to adapt to new and different kinds of bots.

This paper proposes a research framework to enhance the generalization of social bot detection from two dimensions: feature extraction and detection approaches. First, 36 features are extracted from four views for social bot detection. Then, this paper analyzes the feature contribution in different kinds of social bots, and the features with stronger generalization are proposed. Finally, this paper introduces outlier detection approaches to enhance the ever-changing social bot detection.

The experimental results show that the more important features can be more effectively generalized to different social bot detection tasks. Compared with the traditional binary-class classifier, the proposed outlier detection approaches can better adapt to the ever-changing social bots with a performance of 89.23 per cent measured using the F1 score.

Based on the visual interpretation of the feature contribution, the features with stronger generalization in different detection tasks are found. The outlier detection approaches are first introduced to enhance the detection of ever-changing social bots.

The purpose of this article was to confirm whether users’ interests are reflected by tweeted Web pages, and to evaluate the credibility of interest reflection of tweeted Web pages.

Interest reflection of Twitter is investigated based on the context of sharing behavior. A context-oriented approach is proposed to evaluate the interest reflection of tweeted Web pages based on machine learning. Some different distribution models of similarity are present, and infer whether tweeted Web pages reflect respective users’ interests by analyzing user access profiles.

The analysis of browsing behaviors finds that many users partially hide their own concerns, hobbies and interests, and emphasize the concerns about social phenomenon. The extensive experimental results showed the context-oriented approach is effective on real net view data.

As the first-of-its-kind study on evaluating the credibility of interest reflection on Twitter, extensive experiments have been conducted on the data sets containing real net view data. For higher accuracy and less subjectivity, various features are generated from user’s Web view and Twitter submission background with some different context factors.