Analyzing Information Diffusion Process over Social Networks
Hiroshi Motoda
Professor
Osaka University and AFOSR/AOARD, Japan
Abstract :
The rise of the Internet and the World Wide Web accelerates the creation of various large-scale social networks, where nodes (vertices) correspond to people or some social entities, and links (edges) correspond to social interaction between them. Clearly these social networks reflect complex social structures and distributed social trends. A social network can also play an important role as a medium for the spread of various types of information in the form of so-called ``word-of-mouth'' communications. For example, innovation, hot topics and even malicious rumors can propagate through social networks, and computer viruses can diffuse through email networks. Widely-used information diffusion models are not deterministic but probabilistic. Because of this probabilistic nature, influence of a node over a network is defined in terms of its expectation (average number of nodes it influences). Thus, analyzing how information diffuses over social networks involves heavy computation. I first show that a typical social network is different from a random network and address the difficulty associated with the information diffusion problem. Then I explain the networks (blog and Wikipedia) and the information diffusion models (independent cascade, linear threshold and their variants) we used for our analyses.
Bio:
Hiroshi Motoda is scientific advisor of AFOSR/AOARD (Asian Office of Aerospace Research and Development, Air Force Office of Scientific Research, US Air Force Research Laboratory) and professor emeritus of Osaka University, guest professor of the Institute of Scientific and Industrial Research of Osaka University, and visiting professor of the school of computing of University of New South Wales.