|
Abstract:
|
Analyzing social media to combat dark networks, like terrorist networks, provides a number of challenges. On one hand, attribute information associated with social media accounts that support ISIL might be lacking attribute information or deliberately be falsifying that information. On the other hand, social media data provides such a large amount of relational information that incorporating all of the relational information of about an account of interest, and all of the accounts that it is associated with, can take a very long time. This paper seeks to demonstrate how a random sampling of a social media account's connections can help improve our knowledge about the original account of interest, as well as other possible accounts of interest.
|
