The Analysis and Modeling of Large Linked Networks
NSF Award 0514429
PI:
John
Hopcroft, Cornell University
Co-PI:
Bart Selman, Cornell University
Grad Students: Lukas Kroc, current PhD
student
Daniel Sheldon,
current PhD student
Research Summary Web Spam
References
Sucheta Soundarajan, current PhD student (associated)
Undergrad
Students:
Anand Bhaskar
Research Summary
Aaron Sidford
Research Summary
Alex Tsiatas
Research Summary
Abstract:
T
his proposal focuses on extracting information from large network structures particularly on closing the gap between theory and empirical observations. The proposal focuses on three aspects of this problem. First, the graph structure of networks tends to be sparse and therefore the previous theory on spectral clustering does not apply. We pursue an analysis of spectral methods for low density graphs. Second, we will extend our earlier work on tracking evolving structure in networks to identify structure previously under represented in results. Third, we will develop models of networks that more realistically capture aspects of networks that are vital to closing the gap between theory and empirical observations.Statement concerning broader impacts resulting from proposed activities: Our work will provide better models and tools for the rapidly growing scientific community studying large networks. Some of the key infrastructure in our modern society relies on large interconnected networks, such as the powergrid and the World Wide Web. A better understanding both in theoretical and practical terms of large networks should enable us to improve the robustness, reliability, and efficiency of such networks.
Anirban Dasgupta, John Hopcroft, Ravi Kannan, and Pradiptra Mitra, "Spectral Clustering by Recursive Partitioning", ESA, vol. , (2006), p. 298. Published