Identifying trolls and determining terror awareness level in social networks using a scalable framework

Mutlu, Busra; Mutlu, Merve; Öztoprak, Kasım; Doğdu, Erdoğan

doi:10.1109/bigdata.2016.7840796

Cited by 15 publications

(7 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since intent is difficult to ascertain, we only coded tweets as disinformation if they were associated with troll accounts, or accounts that used popular memes to distort general perceptions about an event (Mutlu et al, 2016). We began by identifying memes about school shootings using the following links: www.knowyourmeme.com and www.me.me.com, and then searched the (re)tweets in our sample for keywords associated with popular memes such as “Sam Hyde is the shooter,” which emerged after the University of California (UC) Santa Barbara shooting.…”

Section: Cases and Methodsmentioning

confidence: 99%

“…Consequently, if an account was still active and the user retweeted a meme from a troll account, we assumed that the user was unfamiliar with the meme and coded the retweet as misinformation. This categorization is more conservative than that used by other scholars, who label trolls as individuals circulating “malicious” information or “disruptive activity” designed to lure people into pointless debate (Abril, 2016; Mutlu et al, 2016). We used a more conservative approach in an effort to better assess who served as opinion leaders and the quality of information they shared.…”

Section: Cases and Methodsmentioning

confidence: 99%

See 1 more Smart Citation

From “thank god for helping this person” to “libtards really jumped the shark”: Opinion leaders and (in)civility in the wake of school shootings

Rohlinger

Williams

Teek

2019

New Media & Society

View full text Add to dashboard Cite

Drawing on a qualitative analysis of 5996 tweets and 480 mainstream news stories about the Florida State University (FSU) and the Ohio State University (OSU) shootings, we examine who emerges as opinion leaders during crises, the kinds of narratives they help construct about school shootings, and the relative civility of these narratives. We find that the opinion leaders who emerge after a crisis are assumed to have local knowledge about the incident and/or are able to quickly curate information about the incident. In addition, we find that the quality of information spread by opinion leaders is critical to narrative construction and civility. The largely fact-based narratives associated with the FSU incident were far more civil than the OSU narratives, which were based on disinformation and polemics. We conclude the article by calling on scholars to take a more nuanced approach to conceptualizing and studying opinion leaders.

show abstract

Section: Cases and Methodsmentioning

confidence: 99%

Section: Cases and Methodsmentioning

confidence: 99%

From “thank god for helping this person” to “libtards really jumped the shark”: Opinion leaders and (in)civility in the wake of school shootings

Rohlinger

Williams

Teek

2019

New Media & Society

View full text Add to dashboard Cite

show abstract

“…Peimen and others [22] use external lexicons and TF-IDF features extraction for finding a correlation between Twitter sentiment and events that have occurred. Bursa et al [23] identify malicious users on social media who propagate false opinions and distort the general perception, using KNN, Naïve Bayes, and decision tree machine learning algorithms. They also determine terror awareness level on social media using a scalable framework.…”

Section: Related Workmentioning

confidence: 99%

Adaptable Reduced-Complexity Approach Based on State Vector Machine for Identification of Criminal Activists on Social Media

et al. 2021

View full text Add to dashboard Cite

In today's world, the security agencies has an emerging challenge to identify and counter the malicious contents spread on the social media by the terrorists. However, the text classification techniques have limitations like text visualization, pre-processing, features extraction, and larger features space. The objectives of the criminal users also change with time causing the learning models with additional tasking of identifying the altered malicious textual contents. In this paper, a simplified yet adaptable framework is proposed that makes use of a novel features extraction algorithm (FEA) for extracting features from the textual part of social media contents. It extracts a Boolean feature vector of only 8 dimensions, from each textual content in six steps. The first two values of each feature vector contains information about the presence of URL of malicious contents in the text, the next two values convey their hashtag campaign, 5-6th and 7-8th values are related to their sympathetic and hate campaigns respectively. The extracted features are suitably used to train the state vector machine for the classification of the malicious content promulgated by the criminals. The performance of the proposed method is evaluated against other popular feature selection/ extraction algorithms like TF-IDF Information Gain (IG), Gini Index (GI), Chi Square statistics and PCA and machine learning classifiers like Decision Tree (DT), Random Forest (RF), and Naïve Bayes (NB). It is found that the proposed approach consumes less energy on text visualization, pre-processing, and dimensionality reduction. It also reduces the time-space complexity of the features extraction process and is capable to steer according to the changing future strategies of the active criminal groups.The proposed approach is found to effectively analyze the propaganda material publish by the extremists. It automatically identifies the radical text on social media platforms allowing understanding of the behaviors characteristics and subsequent blockage of such content.

show abstract

“…As future work, our next step will be a semantic analysis [38,41], and ranking [18] of the collected SPARQL endpoints. This will help to get a better understanding about the contents of the discovered Linked Data sources, making it possible to classify SPARQL endpoints according to their domain or context [31][32][33]. Moreover, since it was observed that a small number of endpoints was not discovered by SpEnD (due to search engines' limitations) but were still available on the static repositories, we will put some effort into improving our overall approach to better include, in an automatic way, endpoints and datasets listed only on those repositories.…”

Section: Comparison Of Spend With Existing Sparql Endpoints Repositoriesmentioning

confidence: 99%

A Discovery and Analysis Engine for Semantic Web

Yumuşak

Kamilaris

Doğdu

et al. 2018

Companion of the the Web Conference 2018 on the Web Conference 2018 - WWW '18

Self Cite

View full text Add to dashboard Cite

The Semantic Web promotes common data formats and exchange protocols on the web towards better interoperability among systems and machines. Although Semantic Web technologies are being used to semantically annotate data and resources for easier reuse, the ad hoc discovery of these data sources remains an open issue. Popular Semantic Web endpoint repositories such as SPARQLES, Linking Open Data Project (LOD Cloud), and LODStats do not include recently published datasets and are not updated frequently by the publishers. Hence, there is a need for a web-based dynamic search engine that discovers these endpoints and datasets at frequent intervals. To address this need, a novel web meta-crawling method is proposed for discovering Linked Data sources on the Web. We implemented the method in a prototype system named SPARQL Endpoints Discovery (SpEnD). In this paper, we describe the design and implementation of SpEnD, together with an analysis and evaluation of its operation, in comparison to the aforementioned static endpoint repositories in terms of time performance, availability, and size. Findings indicate that SpEnD outperforms existing Linked Data resource discovery methods. CCS CONCEPTS • Information systems → Web searching and information discovery;

show abstract

Identifying trolls and determining terror awareness level in social networks using a scalable framework

Cited by 15 publications

References 8 publications

From “thank god for helping this person” to “libtards really jumped the shark”: Opinion leaders and (in)civility in the wake of school shootings

From “thank god for helping this person” to “libtards really jumped the shark”: Opinion leaders and (in)civility in the wake of school shootings

Adaptable Reduced-Complexity Approach Based on State Vector Machine for Identification of Criminal Activists on Social Media

A Discovery and Analysis Engine for Semantic Web

Contact Info

Product

Resources

About