Social spam detection

Markines, Benjamin; Cattuto, Ciro; Menczer, Filippo

doi:10.1145/1531914.1531924

Cited by 150 publications

(97 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…LibSVM and rotation forest classifiers have slightly lower accuracy of 0.986 and 0.981, with 0.014 and 0.019 of FPR, respectively. As noted by Markines et al [14], in a deployed social spam detection system it is more important that FPR is kept low compared to high accuracy, because misclassification of a legitimate user is a more consequential mistake than missing a spammer. Other researchers, who proposed different features from the whole or partial dataset of ECML PKDD Discovery Challenge 2008, obtained similar results, for example, Markines et al [14] were able to reach 0.979 of accuracy and 0.013 of FPR, while Bogers et al [3] got 0.9799 of classification accuracy.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Spam Fighting in Social Tagging Systems

Yazdani

Ivanov

Analoui

et al. 2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Tagging in online social networks is very popular these days, as it facilitates search and retrieval of diverse resources available online. However, noisy and spam annotations often make it difficult to perform an efficient search. Users may make mistakes in tagging and irrelevant tags and resources may be maliciously added for advertisement or self-promotion. Since filtering spam annotations and spammers is time-consuming if it is done manually, machine learning approaches can be employed to facilitate this process. In this paper, we propose and analyze a set of distinct features based on user behavior in tagging and tags popularity to distinguish between legitimate users and spammers. The effectiveness of the proposed features is demonstrated through a set of experiments on a dataset of social bookmarks.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Markines et al [14] proposed six different tag-, content-and user-based features for automatic detection of spammers in BibSonomy. The authors used features representing the probability of a tag being spam, number of advertises per post and number of valid resources per user posts.…”

Section: Spam Fighting In Social Tagging Systemsmentioning

confidence: 99%

Spam Fighting in Social Tagging Systems

Yazdani

Ivanov

Analoui

et al. 2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…The honey profiles are normal user profiles created to passively trap spammers by tracking down their malicious activities towards the profiles. Markines et al (2009) explored the six features of social spam appeared on social tagging system, including TagSpam, TagBlur, DomFp, NumAds, Plagiarism, and Validlinks. The six explored features can be well fitted in various machine learning algorithms for classification.…”

Section: Malicious Actor Detecting Mechanisms For Communitiesmentioning

confidence: 99%

A computational framework for detecting malicious actors in communities

Zhan

Fang

2014

IJIPSI

View full text Add to dashboard Cite

Despite the significant research achievements on the study of communities, how to maintain a benign social environment for a community, as a problem, has not received much attention. Current existing malicious activity detecting mechanisms are subject to the limitation of the underlying online environment. However, we found that information plays an important role in terms of socialisation. Malicious activities, like diffusing unreliable information and providing inappropriate critics, can severely impact the situation of a community's social environment. In this paper, we propose a computational framework for detecting malicious actors in communities from the perspective of information diffusion. We use the term, 'malicious actors', to represent a group of people who intentionally or unintentionally conduct malicious behaviours to sabotage benign social environments of communities.Keywords: malicious actors; information diffusion; computational framework.Reference to this paper should be made as follows: Zhan, J. and Fang, X. (2014) 'A computational framework for detecting malicious actors in communities', Int.

show abstract

“…In a social tagging system, spam or noise can be injected at three different levels: spam content, spam tag-content association, and spammer [18]. Trust modeling can be performed at each level separately (e.g., [18]) or different levels can be considered jointly to produce trust models, for example, to assess a user's reliability, one can consider not only the user profile, but also the content that the user uploaded to a social system (e.g., [19]).…”

Section: Trust Modelingmentioning

confidence: 99%

“…Trust modeling can be performed at each level separately (e.g., [18]) or different levels can be considered jointly to produce trust models, for example, to assess a user's reliability, one can consider not only the user profile, but also the content that the user uploaded to a social system (e.g., [19]). In this article, we categorize trust modeling approaches into two classes according to the target of trust, i.e., user and content trust modeling (shown in Figure 3).…”

Section: Trust Modelingmentioning

confidence: 99%