Mutual clustering coefficient-based suspicious-link detection approach for online social networks

Understanding of social network structure and user behavior has important implications for site design, applications (e.g., ad placement policies), accurate modeling for social studies, and design of next-generation infrastructure and content distribution systems. Currently, characterizations of social networks have been dominated by topological studies in which graph representations are analyzed in terms of connectivity using techniques such as degree distribution, diameter, average degree, clustering coefficient, average path length, and cycles. The problem is that these parameters are not completely satisfactory in the sense that they cannot account for individual events and have only limited use, since one can produce a set of synthetic graphs that have the exact same metrics or statistics but exhibit fundamentally different connectivity structures. In such an approach, a node drawn as a small circle represents an individual. A small circle reflects a black box model in which the interior of the node is blocked from view. This paper focuses on the node level by considering the structural interiority of a node to provide a more fine-grained understanding of social networks. Node interiors are modeled by use of six generic stages: creation, release, transfer, arrival, acceptance, and processing of the artifacts that flow among and within nodes. The resulting description portrays nodes as comprising mostly creators (e.g., of data), receivers/senders (e.g., bus boys), and processors (re-formatters). Two sample online social networks are analyzed according to these features of nodes. This examination points to the viability of the representational method for characterization of social networks.

Download Full-text

EFFECTIVENESS OF OPINION INFLUENCE APPROACHES IN HIGHLY CLUSTERED ONLINE SOCIAL NETWORKS

Advances in Complex Systems ◽

10.1142/s0219525914500088 ◽

2014 ◽

Vol 17 (02) ◽

pp. 1450008

Author(s):

MELISSA FALETRA ◽

NATHAN PALMER ◽

JEFFREY S. MARSHALL

Keyword(s):

Social Networks ◽

Social Network ◽

Online Social Networks ◽

Online Social Network ◽

Threshold Value ◽

Clustering Coefficient ◽

Scale Free ◽

Small Influence ◽

Online Advertisement ◽

The Difference

A mathematical model was developed for opinion propagation on online social networks using a scale-free network with an adjustable clustering coefficient. Connected nodes influence each other when the difference between their opinion values is less than a threshold value. The model is used to examine effectiveness of three different approaches for influencing public opinion. The approaches examined include (1) a "Class", defined as an approach (such as a class or book) that greatly influences a small, randomly selected portion of the population, (2) an "Advertisement", defined as an approach (such as a TV or online advertisement) that has a small influence at each viewing on a large randomly selected portion of the population, and (3) an "App", defined as an approach (such as a Facebook game or smartphone "App") that spreads via the online social network (rather than randomly) and has a small influence at each viewing on the affected population. The Class and Advertisement approaches result in similar overall influence on the population, despite the fact that these approaches are highly different. In contrast, the App approach has a much more significant effect on opinion values of users occupying clusters within the social network compared to the overall population.

Download Full-text

Unsupervised Spammer Detection in Social Networks based on User Information Interests

10.21203/rs.3.rs-798887/v1 ◽

2021 ◽

Author(s):

Darshika Koggalahewa ◽

Yue Xu ◽

Ernest Foo

Keyword(s):

Social Networks ◽

Social Network ◽

Online Social Networks ◽

Supervised Classification ◽

Peer Acceptance ◽

Spam Detection ◽

Highly Active ◽

Detection Approach ◽

Data Fabrication ◽

Unsupervised Approach

Abstract Online Social Networks (OSNs) are a popular platform for communication and collaboration. Spammers are highly active in OSNs. Uncovering spammers has become one of the most challenging problems in OSNs. Classification-based supervised approaches are the most commonly used method for detecting spammers. The classification-based systems suffer from limitations of “data labelling”, “spam drift”, “imbalanced datasets” and “data fabrication”. These limitations effect the accuracy of a classifier’s detection. We present a pure unsupervised approach for spammer detection based on peer acceptance of a user in a social network to distinguish spammers from genuine users. The peer acceptance of a user to another user is calculated based on common shared interests over multiple shared topics between the two users. The main contribution of this paper is the introduction of a pure unsupervised spammer detection approach based on users’ peer acceptance. Our approach does not require labelled training datasets. While it does not better the accuracy of supervised classification-based approaches, our approach has become a successful alternative for traditional classifiers for spam detection by achieving an accuracy of 96.9%.

Download Full-text

An unsupervised method for social network spammer detection based on user information interests

Journal Of Big Data ◽

10.1186/s40537-021-00552-5 ◽

2022 ◽

Vol 9 (1) ◽

Author(s):

Darshika Koggalahewa ◽

Yue Xu ◽

Ernest Foo

Keyword(s):

Social Networks ◽

Social Network ◽

Online Social Networks ◽

Supervised Classification ◽

Peer Acceptance ◽

Spam Detection ◽

Highly Active ◽

Detection Approach ◽

Data Fabrication ◽

Unsupervised Approach

AbstractOnline Social Networks (OSNs) are a popular platform for communication and collaboration. Spammers are highly active in OSNs. Uncovering spammers has become one of the most challenging problems in OSNs. Classification-based supervised approaches are the most commonly used method for detecting spammers. Classification-based systems suffer from limitations of “data labelling”, “spam drift”, “imbalanced datasets” and “data fabrication”. These limitations effect the accuracy of a classifier’s detection. An unsupervised approach does not require labelled datasets. We aim to address the limitation of data labelling and spam drifting through an unsupervised approach.We present a pure unsupervised approach for spammer detection based on the peer acceptance of a user in a social network to distinguish spammers from genuine users. The peer acceptance of a user to another user is calculated based on common shared interests over multiple shared topics between the two users. The main contribution of this paper is the introduction of a pure unsupervised spammer detection approach based on users’ peer acceptance. Our approach does not require labelled training datasets. While it does not better the accuracy of supervised classification-based approaches, our approach has become a successful alternative for traditional classifiers for spam detection by achieving an accuracy of 96.9%.

Download Full-text