Social circle analysis on ego-network based on context frequent pattern mining

Abstract

Effective organization and management of a user's ego-network is one of the effective means to deal with information overload. Manual organization methods are often time-consuming and difficult to promote. In this paper, an automatic identification of social circles model is proposed, which takes three factors into account: user's profile, relationship information, and interactive context information. Using closed frequent itemsets mining on users' interactive context, a user in egonetwork can belong to more than one social circles or even no one. The experiments show that our model can effectively detect social circles on ego-network.

Predicting the dynamics of social circles in ego networks using pattern analysis and GA K‐means clustering

“…One fundamental assumption of the model is that circles will be made up of close knit group friends who also share common properties or traits, but disregards the interactive contextual information of users. Xu et al, further studied the problem and consider not only the impact of individual attributes, but also care about the influence of interaction between individuals during circle formation. The similar problem is also studied by MacLean et al, but in the context of Email.…”

“…A number of participants in a survey reported that they were unaware of such functionality exists, or found it cumbersome or confusing to configure lists or customize privacy settings based on these lists . Thus, finding automated means of creating circles to provide semantically meaningful grouping of connections is now becoming a challenging area . It also provides researchers a rich and unprecedented opportunity to the analysis and model relationships among various objects in social networks with the aim of understanding how the behavior of individuals and their interactions translates into large‐scale social phenomenon .…”

“…12 Thus, finding automated means of creating circles to provide semantically meaningful grouping of connections is now becoming a challenging area. 8,[13][14][15][16][17][18] It also provides researchers a rich and unprecedented opportunity to the analysis and model relationships among various objects in social networks with the aim of understanding how the behavior of individuals and their interactions translates into large-scale social phenomenon. 19,20 Traditionally, the aim of community detection algorithms in the graph is to identify the modules and, possibly, their hierarchical organization, by only using the information encoded in the graph topology.…”

“…In other words, it uses an unsupervised method to learn what is/are the dimension(s) of users' profile similarity along which each circle emerges. Xu et al have also studied the problem of organizing and managing a user's ego network not only by considering user's profile and relationship information, but also the frequent sets of items appear during interactions between users. The problem of automatic discovery of social groups on Email has also been studied by MacLean et al To address this problem, they initially create social topologies, a set of potentially overlapping and nested social groups , and identify likely groups based on co‐occurrence patterns.…”