Faloutsos KDD 2005

From Cohen Courses
Jump to navigationJump to search

Citation

J. Leskovec, J. Kleinberg and C. Faloutsos. Graphs over Time: Densification Laws, Shrinking Diameters and Possible Explanations. ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2005.

Online version

KDD 2005

Summary

This paper presents an approach to Graph Generation. Here they try to answer the question of network evolution of over time. While static graph models have been studied, less is known about time-evolving graphs. They use the following Patent Citation Network. They study a number of real world graphs and observe some interesting phenomena.

  • Most graphs densify over time with number of edges growing superlinearly with respect to the number of nodes
  • The average distance between nodes shrinks over time

Since existing graph generation models do not model for such behaviors they propose a new graph generation model based on a forestfire approach.

Method

They use the Forest Fire method for graph generation. The goal of this method to construct a graph that has the properties of following densification power laws, long-tailed in and out degrees and shrinking diameter.