Mining of Leaders in Mobile Telecom Social Networks

Mining of Leaders in Mobile Telecom Social Networks

Mantian (Mandy) Hu (The Chinese University of Hong Kong, China)
Copyright: © 2017 |Pages: 10
DOI: 10.4018/978-1-5225-1750-4.ch006
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

In the age of Big Data, the social network data collected by telecom operators are growing exponentially. How to exploit these data and mine value from them is an important issue. In this article, an accurate marketing strategy based on social network is proposed. The strategy intends to help telecom operators to improve their marketing efficiency. This method is based on mutual peers' influence in social network, by identifying the influential users (leaders). These users can promote the information diffusion prominently. A precise marketing is realized by taking advantage of the user's influence. Data were collected from China Mobile and analyzed. For the massive datasets, the Apache Spark was chosen for its good scalability, effectiveness and efficiency. The result shows a great increase of the telecom traffic, compared with the result without leader identification.
Chapter Preview
Top

Social Network Behavior

In sociology, an individual’s behavior is not only determined by his character, but also by his friends’ behavior. Then, when we analyze an individual behavior, we need to consider his friends effect. Godinho de Matos, Ferreira, and Krackhardt (2014) estimate peer influence in the diffusion of iPhone 3G across different communities. They find that if one’s friend has iPhone 3G, there is a greater probability for him to switch to this cell phone when changing it. Pushpa (2012) construct a telecom social network, and find that the efficiency to predict customer churn based on social network has a good increase.

In communication networks, users build relationships each other by mobile communication like calls and messages. This relationship between individuals can been seen as a graph, each node of a graph representing a user. An example of such graph is depicted on Figure 1.

Figure 1.

Communication network example

Often, if the original network is large, it is necessary to construct or identify subpopulations from this large network, before using them for analysis. In complex networks, there are many ways to detect a community. Newman (2004) proposes a modularity spectral optimization, uses modularity as an evaluation criteria and detects subpopulations by maximization modularity. Clauset, Newman, and Moore (2005) design a greedy maximization of the local modularity algorithm. This algorithm is considered to be one of the fastest community extraction algorithms. For the communication circle’s character, we use the snowball sampling method to extract the subgraph from the whole mobile users.

In statistics research, snowball sampling (Goodman, 1961) is a non-probabilistic sampling technique and is a useful tool for building networks and increasing the number of participants. We use a snowball sampling method to get the seed user’s contacts, and his contacts’ contacts. Therefore, we obtain the seed user’s two-layers contacts using this sampling method. An example of such two-layers contacts is shown on Figure 2.

Figure 2.

Two-layers snowball sampling of a user

After having performed community detection, we use the Internal-External Ratio index (Zhang, 2011) to optimize this community, eliminating the low coherent users. Consequently, users in this circle are closer to each other and relationship between different circle becomes sparse.

Complete Chapter List

Search this Book:
Reset