Marketing on Dynamical Random Networks and Related Inference

• Activity Number: 184
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: IMS
• Author(s): Daniel Saxton*+ and Anand Vidyashankar
• Companies: and George Mason University
• Keywords: scale-free networks ; preferential attachment ; fitness ; hub ; power-law

Abstract

Statistical methods for network data is an emerging field with intense focus from several statisticians. However, statistical evaluation of features of dynamical networks and their influence on marketing strategies is an important and challenging problem that has not been investigated. In this talk, we will provide several probabilistic models for describing cascade dynamics and statistical methods to evaluate the parameters of the cascade processes and the networks. Specifically, we will show that in a variety of models used for describing cascade dynamics, the degree distribution of the network exhibits a power-law. We will also characterize the growth of the degrees of nodes. Using these results, we develop a principled approach to estimating the node influence, predicting the link structure, and identifying the community structure within the network.