Predicting Consumer Choices Through Analysis of Interactions in Social Networks

Abstract

Analysis of interactions in social networks has emerged as a new research paradigm in modern marketing. It focuses not on modeling behavior of the individual but rather than on the measurement and analysis of its relationships and interactions with other users within the network. Measurement and analysis of these interactions can help understand the structure and dynamics of social networks and their impact on consumer choice. In this paper we present a data mining approach to measure and analyze the interactions in social networks between clients of mobile telecommunication networks. Our goal is to demonstrate how to use Call Data Records (CDR) to build predictive choice models (e.g. to predict customer churn). The approach and methodology can be applied to analyze the customer choice behavior in other markets where customer interactions are tracked automatically and saved electronically.