Die Free or Live Hard? Empirical Evaluation and New Design for Fighting Evolving Twitter Spammers

Yang, Chao; Harkreader, Robert Chandler; Gu, Guofei

doi:10.1007/978-3-642-23644-0_17

Cited by 182 publications

(144 citation statements)

References 15 publications

Supporting

Mentioning

140

Contrasting

Unclassified

Order By: Relevance

“…Some systems analyze the characteristics of social network accounts, looking for sign of mass-created fake accounts [7,16,24,31,40,42]. Other systems look at the social network structure of accounts, or at how the messages posted by them spread, looking for anomalies indicative of a malicious profile [8,9,15,32,38,41].…”

Section: Related Workmentioning

confidence: 99%

Quit Playing Games with My Heart: Understanding Online Dating Scams

Huang

Stringhini

Yong

2015

Detection of Intrusions and Malware, and Vulnerability Assessment

View full text Add to dashboard Cite

Abstract. Online dating sites are experiencing a rise in popularity, with one in five relationships in the United States starting on one of these sites. Online dating sites provide a valuable platform not only for single people trying to meet a life partner, but also for cybercriminals, who see in people looking for love easy victims for scams. Such scams span from schemes similar to traditional advertisement of illicit services or goods (i.e., spam) to advanced schemes, in which the victim starts a long-distance relationship with the scammer and is eventually extorted money. In this paper we perform the first large-scale study of online dating scams. We analyze the scam accounts detected on a popular online dating site over a period of eleven months, and provide a taxonomy of the different types of scammers that are active in the online dating landscape. We show that different types of scammers target a different demographics on the site, and therefore set up accounts with different characteristics. Our results shed light on the threats associated to online dating scams, and can help researchers and practitioners in developing effective countermeasures to fight them.

show abstract

Section: Related Workmentioning

confidence: 99%

Quit Playing Games with My Heart: Understanding Online Dating Scams

Huang

Stringhini

Yong

2015

Detection of Intrusions and Malware, and Vulnerability Assessment

View full text Add to dashboard Cite

show abstract

“…While the TPR is similar to those seen with other algorithms [6,2,8,10], direct comparison is difficult due to variations in methodologies. In particular, differences in determining a suitable "ground truth" (Twitter suspension information, URL blacklists, and manual verification) and granularity (account and message levels) mean each study is measuring a slightly different value.…”

Section: Resultsmentioning

confidence: 61%

“…Others techniques use both the content of the messages, profile information, and information from the Twitter social graph to try and determine the nature of tweets [6,7,8,9,10,11]. Even more complex methods have further used similar types of information to determine which large scale campaign a spam tweet belongs [2,4,12,3].…”

Section: Introductionmentioning

confidence: 99%

Searching for Spam: Detecting Fraudulent Accounts via Web Search

Flores¹,

Kuzmanovic²

2013

Passive and Active Measurement

View full text Add to dashboard Cite

Abstract. Twitter users are harassed increasingly often by unsolicited messages that waste time and mislead users into clicking nefarious links. While increasingly powerful methods have been designed to detect spam, many depend on complex methods that require training and analyzing message content. While many of these systems are fast, implementing them in real time could present numerous challenges.Previous work has shown that large portions of spam originate from fraudulent accounts. We therefore propose a system which uses web searches to determine if a given account is fraudulent. The system uses the web searches to measure the online presence of a user and labels accounts with insufficient web presence to likely be fraudulent. Using our system on a collection of actual Twitter messages, we are able to achieve a true positive rate over 74% and a false positive rate below 11%, a detection rate comparable to those achieved by more expensive methods.Given its ability to operate before an account has produced a single tweet, we propose that our system could be used most effectively by combining it with slower more expensive machine learning methods as a first line of defense, alerting the system of fraudulent accounts before they have an opportunity to inject any spam into the ecosystem.

show abstract

“…When the user selects the name of an existing anti spammer system, WEST automatically selects the set of features used by that system. WEST also allows for new features to be added by the user and presently there are 17 systems [1,2,4,[6][7][8][9][10][11][12][13][14][15][16][17][18][19] and 173 features. Each of those features can be grouped into four types: profile, activities, relations, and tweet contents.…”

Section: Feature Extractionmentioning

confidence: 99%

“…They deploy different techniques to post unwanted messages to users on SNS for advertisement, frauds or spreading of malware through the malicious URLs [12]. For instance, spammers create many fake accounts to post spam tweets for a specific purpose (or known as spam campaign), send message with different text to convey the same meanings or pay some users to follow their accounts [15]. Thus, the statistical attributes of spammers and spam tweets vary over time.…”

Section: Introductionmentioning

confidence: 99%

A Framework for Evaluating Anti Spammer Systems for Twitter

Liesaputra

Yongchareon

et al. 2017

On the Move to Meaningful Internet Systems. OTM 2017 Conferences

View full text Add to dashboard Cite

Abstract. Despite several benefits to modern communities and businesses, Twitter has attracted many spammers overwhelming legitimate users with unwanted and disruptive advertising and fake information. Detecting spammers is always challenging because there is a huge volume of data that needs to be analyzed while at the mean time spammers continue learning and changing their ways to avoid being detected by anti-spammer systems. Several spam classification systems are proposed using various features extracted from the content and user's information from their Tweets. Nevertheless, no comprehensive study has been done to compare and evaluate the effectiveness and efficiency of these systems. It is not known what the best anti-spammer system is and why. This paper proposes an evaluation framework that allows researchers, developers, and practitioners to access existing user-based and content-based features, implement their own features, and evaluate the performance of their systems against other systems. Our framework helps identify the most effective and efficient spammer detection features, evaluate the impact of using different numbers of recent tweets, and therefore obtaining a faster and more accurate classifier model. Keywords: Spam detection; Evaluation workbench; Feature selection; Machine learning IntroductionSpams are unwanted activities such as when marketers send members unwanted advertisements, post fake reviews, or steal user information by directing users to malicious external pages [11]. As Social Network Services (SNS) becoming an important mode of communication, it attracts spammers who overwhelms users with unwanted content. Among these sites, Twitter, which was started in 2006, has grown to be one of the most popular SNS [22]. There are 500 million number of messages (called tweets) produced by 328 million active Twitter users (called twitterers) every day. Unlike other popular SNS, tweets can be read by anyone and people can follow a user without their consent. To attract users to their target websites, spammers post a large 2 number of coordinated messages containing specific URLs and sometimes describing them with unrelated words [26]. Because SNS helps build intrinsic trust between their users, 45% of them will click on links posted by their online friends even though they do not know those people in real life [24]. Twitterers also tend to post shortened URLs and write in abbreviated forms that rarely appear in conventional text documents or e-mails as a tweet can only contain up to 140 characters. Consequently, it is difficult for users to know the source URL and identify the content of the URL without clicking the link and loading the page. The noisy, unstructured, and informal expressions, such as "2mo is a new daaaaay!" or "TIL DC Comics stands for Detective Comics", used in the text also made it difficult for automatic spam detection system to accurately identify the semantic meaning of the tweets. Hence, social spamming is more harmful and complex than SMS, email or Web spams. It is bec...

show abstract

Die Free or Live Hard? Empirical Evaluation and New Design for Fighting Evolving Twitter Spammers

Cited by 182 publications

References 15 publications

Quit Playing Games with My Heart: Understanding Online Dating Scams

Quit Playing Games with My Heart: Understanding Online Dating Scams

Searching for Spam: Detecting Fraudulent Accounts via Web Search

A Framework for Evaluating Anti Spammer Systems for Twitter

Contact Info

Product

Resources

About