Structure and Dynamics of Social Networks and Other Networked Systems

点击翻译按钮获取中文摘要

Newman0109086 The principal investigator and his colleagues study thestructure and function of real-world networked systems,particularly but not exclusively social networks. An empiricalcomponent is concerned with the discovery and analysis of thestructure of networks, including networks of collaborationbetween scientists, networks of company directors, networks ofpersonal preferences, and networks of citations between academicpublications. Studied quantities include local observables suchas transitivity and degree distribution, and nonlocal ones suchas centrality and community structure. The investigator developsmodels to aid in the understanding of the effects of networkstructure. Of particular interest are random graph models ofnetworks, percolation models of network resilience, and models ofepidemics taking place on social networks. The investigatordevelops new algorithms for extracting and visualizing networkstructure, particularly the existence of communities in networksand structural properties related to network resilience, such aspath counts and centrality measures. A knowledge of the structure of networks of acquaintance iscrucial to the understanding of how information, such as news,rumors, consumer trends, etc., spreads through society.Similarly, networks of physical contact between people govern theway in which diseases spread. A proper understanding of thenature and progress of epidemics is impossible without goodnetwork models. In this project the investigator determines whatthe structure of the networks in question is, and also models theeffect of that structure on, among other things, the spread ofinformation and disease. As well as enhancing basicunderstanding of these problems, the project points to ways inwhich network structure or dynamics can be changed in order toeither improve network transmission (in the case of information)or slow it down (in the case of epidemics). For diseasetransmission, for instance, it may be able to suggest effectivetargets for immunization or education campaigns to slow diseasespread. The new data resources and analysis techniques developedcan be used to study other problems in which network-structuredprocesses arise.