A Local Structure-Based Method for Nodes Clustering: Application to a Large Mobile Phone Social Network
In this paper we present a method for describing how a node of a given graph is connected to the network. We also propose a method for grouping nodes into clusters based on the structure of the network in which they are embedded, so on the description pro
- PDF / 1,046,035 Bytes
- 28 Pages / 439.36 x 666.15 pts Page_size
- 19 Downloads / 189 Views
A Local Structure-Based Method for Nodes Clustering: Application to a Large Mobile Phone Social Network Alina Stoica, Zbigniew Smoreda, and Christophe Prieur
Abstract In this paper we present a method for describing how a node of a given graph is connected to the network. We also propose a method for grouping nodes into clusters based on the structure of the network in which they are embedded, so on the description provided by the first method. We apply these methods to a mobile phone communications network. When confronting the obtained clusters of individuals to their age and to their intensity of communication, the results are quite promising: the two measures are correlated to the social network cluster. We finish by providing a typology of the mobile phone users based on social network cluster, communication intensity and age.
7.1 Introduction In this paper, we want to describe how each individual of a given social network is connected to the network and to cluster nodes that are connected in a similar way to the network. One can see this distribution of nodes into clusters as an identification of network “roles”. Without pretending to have solved the problem of identification of roles, we present a method to distribute nodes into clusters based on the local structure of the network.
A. Stoica () EDF R&D, 1 av. du Gen. de Gaulle Clamart, Clamart, France e-mail: [email protected] Z. Smoreda Orange Labs, 38-40 rue du Gen. Leclerc, Issy les Moulineaux, France e-mail: [email protected] C. Prieur LIAFA, Paris-Diderot, 75 rue du Chevaleret, 75013 Paris, France e-mail: [email protected] T. Özyer et al. (eds.), The Influence of Technology on Social Network Analysis and Mining, Lecture Notes in Social Networks 6, DOI 10.1007/978-3-7091-1346-2__7, © Springer-Verlag Wien 2013
157
158
A. Stoica et al.
We apply the method to a large mobile phone network. The obtained results are quite promising, in particular when the clusters are confronted to other characteristics of the individuals. Indeed the probability that an individual belongs to a certain cluster depends on his or her age; even more, using these probabilities we are able to group together different ages, thus discovering four groups containing consecutive ages, corresponding to four life stages. The probability that a person belongs to a certain cluster also depends on his or her mobile phone communication intensity; moreover the intensity of communication allows us to predict with rather high accuracy the cluster membership of a person. We begin by recalling some work related to the subject. Next we present the method for describing how a node of a given graph is embedded in the network. We then propose a method for clustering nodes based on the structure of the network surrounding them. Next we present the results of the application of the methods to the mobile phone communications network: the obtained clusters of individuals, the correlations with the age and the intensity of communication and a typology of mobile phone users
Data Loading...
