mickey - Local Algorithms for Finding Interesting...

Info iconThis preview shows pages 1–8. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Local Algorithms for Finding Interesting Individuals in Large Social Networks Mickey Brautbar, Michael Kearns Computer and Information Science University of Pennsylvania December 8, 2009 Structural Properties of Social Networks Nodes with high degree. Nodes with high clustering coefficient. Nodes with high betweenness centrality (characterizes whether a lot of paths must cross through the node at hand). Small diameter. Many other structural properties... Figure: Taken from Complex Graphs and Networks by Chung and Lu. The Algorithmic Side Question 1: If the network contains the structure at hand, how hard is it to find that structure? Question 2: Is it substantially easier to find such a structure in a social network than finding such a structure in an arbitrary network (that contains the structure)? Potential applications. Example: Marketing of new product of small budget companies. Find a node with high degree in an online social network and pay that node to add a link to your new product from his online profile. Guarantees that at least that nodes immediate friends are exposed to the new product. What Kind of Query Model Should We Use? Facebook as a model As a user of Facebook I am lacking a cenrtal view of the Facebook network. The edges in the network are the connections between friends. How hard would it be for me to find a user with lots of friends? I can make a Crawl query - go to the profile of my friend and check his degree. I can make a Jump query by using the Friends Finder option of Facebook. Typing a random name in Friends Finder will return a user with that profile (if one exists). The Model At a given user we know his degree and also the identity of his immediate neighbors. We can make a Crawl query to move to a specified neighbor of the node at hand. Or we can make a Jump query to move to a uniformly at random chosen node from the network. Goals In this talk we shall focus on the high degree structure and also the high clustering coefficient structure....
View Full Document

Page1 / 19

mickey - Local Algorithms for Finding Interesting...

This preview shows document pages 1 - 8. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online