Determining credibility from social network structure

Abstract

The increasing proliferation of social media results in users that are forced to ascertain the truthfulness of information that they encounter from unknown sources using a variety of indicators (e.g. explicit ratings, profile information, etc.). Through human-subject experimentation with an online social network-style platform, our study focuses on the determination of credibility in ego-centric networks based on subjects observing social network properties such as degree centrality and geodesic distance. Using manipulated social network graphs, we find that corroboration and degree centrality are most utilized by subjects as indicators of credibility. We discuss the implications of the use of social network graph structural properties and use principal components analysis to visualize the reduced dimensional space.

Classification of unlabeled online media

“…We assume that users recruited in our social network are not known to each other and by showing every user the credibility of all users, we assume that users tend to follow or be followed by other users on the basis of credibility. We observe a positive linear correlation between the number of links connected to a user (in-degree) and credibility, consistent with findings from other works 27 (for further details see Supplementary Information, section 2). However, in reality, such a credibility measure is often network-dependent, or may not be readily available.…”

“…Though the most direct measure of user credibility involves asking the other users 26 , such a measure is often subjective. Additionally, a subjective measure does not apply to sources or users previously unknown to a user 27 . However, such occasions are likely to occur during times of disasters such as COVID-19 7 , or during measles outbreaks 13,14 where authentic information was provided by those physically attending the event 27 .…”

“…Additionally, a subjective measure does not apply to sources or users previously unknown to a user 27 . However, such occasions are likely to occur during times of disasters such as COVID-19 7 , or during measles outbreaks 13,14 where authentic information was provided by those physically attending the event 27 . Hence, we are interested in understanding how fake information spreads in a controlled social media network where users respond to unknown users with the help of an objective credibility measure.…”

Towards Window Analytics over Large-scale Graphs

“…These neighboring vertices typically represent the most important vertices to a vertex with regard to their structural relationship in a graph. Thus, k-hop windows provide meaningful specifications for many applications, such as customer behavior analysis [3,11] , digital marketing [14] etc.…”

Credibility Ranking Methods Analysis of Users in Social Network Based on Relation Graph

“…Machine learning-based methods can be supervised or unsupervised, and they are based on building classifiers that determine credibility scores for blogs as a measure of their factuality. Examples of this type are presented in [5,11,12]. These methods require recursive processing of data to train and test the algorithms to achieve the desired accuracy.…”